1: %\documentclass[manuscript]{aastex}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\documentclass[12pt,preprint,showpacs,prd,nofootinbib]{revtex4}
5: %\documentclass[12pt,preprint,prd,nofootinbib]{revtex4}
6: \usepackage{graphicx}
7: %\usepackage{psfrag}
8: %\usepackage{amssymb,amsfonts}
9: %\usepackage{showkeys}
10:
11: \pagestyle{headings}
12:
13: \newcommand{\dd}{\mathrm{d}}
14: \newcommand{\ee}{\mathrm{e}}
15: \newcommand{\ii}{\mathrm{i}}
16: \newcommand{\eqref}[1]{(\ref{#1})}
17:
18: \shorttitle{Inelastic Collisions in General Relativity}
19: \shortauthors{Hennig et al.}
20:
21: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22:
23: \begin{document}
24:
25: \title{Thermodynamic Description of Inelastic Collisions in General
26: Relativity}
27:
28: \author{J\"org Hennig, Gernot Neugebauer}
29: \affil{Theoretisch-Physikalisches Institut,
30: Friedrich-Schiller-Universit\"at
31: Jena, Max-Wien-Platz 1, D-07743 Jena, Germany}
32: \email{J.Hennig@tpi.uni-jena.de}
33: \and
34: \author{Marcus Ansorg}
35: \affil{Max-Planck-Institut f\"ur Gravitationsphysik,
36: Albert-Einstein-Institut, Am M\"uhlenberg 1, D-14476 Golm, Germany}
37: \email{marcus.ansorg@aei.mpg.de}
38:
39: %\author{J\"org Hennig\altaffilmark{1}, Gernot Neugebauer\altaffilmark{1}
40: %and Marcus Ansorg\altaffilmark{2}}
41: %\affil{\altaffilmark{1}Theoretisch-Physikalisches Institut,
42: %Friedrich-Schiller-Universit\"at
43: %Jena,\\ Max-Wien-Platz 1, D-07743 Jena, Germany}
44: %\email{J.Hennig@tpi.uni-jena.de}
45: %\affil{\altaffilmark{2}Max-Planck-Institut f\"ur Gravitationsphysik,
46: %Albert-Einstein-Institut,\\ Am M\"uhlenberg 1, D-14476 Golm, Germany}
47: %\email{marcus.ansorg@aei.mpg.de}
48:
49:
50:
51: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
52: \begin{abstract}
53: We discuss head-on collisions of neutron stars and disks of dust (\lq\lq
54: galaxies\rq\rq) following the ideas of equilibrium thermodynamics, which
55: compares equilibrium states and avoids the description of the dynamical
56: transition processes between them. As an always present damping mechanism,
57: gravitational emission results in final equilibrium states after the
58: collision. In this paper we calculate selected final configurations
59: from initial data of colliding stars and disks by making use
60: of conservation laws and solving the Einstein equations. Comparing
61: initial and final states, we can decide for which initial parameters two
62: colliding neutron stars (non-rotating Fermi gas models) merge into a
63: single neutron star and two rigidly rotating disks form again a final
64: (differentially rotating) disk of dust. For the neutron star collision
65: we find a maximal energy loss due to outgoing gravitational radiation of
66: $2.3\%$ of the initial mass while the corresponding efficiency for
67: colliding disks has the
68: much larger limit of $23.8\%$.
69: \end{abstract}
70:
71: \keywords{Equation of state --- gravitation --- gravitational waves ---
72: galaxies: general --- stars: neutron}
73:
74:
75: \maketitle
76:
77: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
78: \section{Introduction}
79:
80:
81: Collisions of compact objects are an important source of gravitational
82: radiation.
83: Much effort has recently been made to develop numerical methods and
84: codes describing and simulating the underlying hydrodynamical and
85: gravitational phenomena.
86: After the pioneering work on numerical black hole
87: evolutions
88: by Eppley and Smarr in the
89: 1970's
90: [see e.g.
91: %\cite{Eppley,Smarr}
92: Eppley (1975) and Smarr et al. (1976)], head-on collisions were re-investigated
93: in the 1990's
94: %\cite{Anninos1, Anninos2, Anninos3}
95: (Anninos et al. 1993, 1995, 1998)
96: with good agreement
97: between numerical and perturbation-theoretical results. Long-term-stable
98: evolutions of black hole and neutron star collisions were
99: successfully performed in the last two years
100: %\cite{Sperhake1, Fiske, Zlochower,Sperhake2, Loeffler}.
101: (Sperhake et al. 2005; Fiske et al. 2005; Zlochower et al. 2005;
102: Sperhake 2006; L\"offler et al. 2006).
103:
104: From a mathematical point of view collision processes are typical
105: examples for initial-boundary problems. In particular,
106: we will discuss head-on
107: collisions of spheres and disks\footnote{Disk-like
108: matter configurations play an
109: important role in astrophysics, e.g. as models for galaxies,
110: accretion disks or intermediate phases in the merger process
111: of two neutron stars.}, see
112: Fig.~\ref{fig_model}.
113: Starting with bodies separated by a large (\lq\lq
114: infinite\rq\rq) distance we may model the initial situation by a
115: quasi-equilibrium configuration of two isolated bodies.
116: Corresponding solutions for spheres and
117: (rigidly rotating) disks can be found in the literature, see e.g.
118: %\cite{MTW,Shapiro,Neugebauer1,Neugebauer2,Neugebauer3}
119: Misner et~al. (2002), Shapiro \& Teukolsky (1983) and Neugebauer \& Meinel
120: (1993, 1994, 1995).
121: The dynamical
122: phase of the collision process is always accompanied by gravitational
123: radiation. This damping mechanism results again in the formation of an
124: equilibrium configuration after the collision.
125: The rigorous mathematical description of the dynamical transition phase
126: is difficult and requires extensive numerical investigations. However,
127: interesting information about the collision can be obtained by comparing
128: the initial and final (equilibrium) states. This thermodynamic idea
129: avoids the analysis of the transition process
130: and reduces the mathematical effort to solving the
131: Einstein equations for the end products, which are stationary and
132: axisymmetric in our case. The
133: solution makes use of conservation laws which transfer
134: data extracted from the initial configurations (spheres and disks
135: before the collision) to the final configurations.
136:
137: While the \emph{initial} configurations are available
138: the calculation of the \emph{final} states is rather difficult.
139: To cope with this problem for head-on colliding stars and disks,
140: we will make use of two
141: heuristic principles:
142: \begin{itemize}
143: \item[1)] Perfect fluid configurations at rest are
144: spherically symmetric. Hence, the end product of two head-on
145: colliding spheres without angular momentum is again a
146: sphere
147: (a fluid ball or a
148: Schwarzschild black hole).
149: \item[2)] Dust configurations
150: are two-dimensional
151: (\lq\lq extremely flattened\rq\rq) and axisymmetric
152: (with non-vanishing angular momentum). Consequently,
153: the dust matter after a head-on
154: collision of two disks of dust is again two-dimensional and
155: axisymmetric (a compact disk, a disk surrounded by
156: dust rings or a black hole surrounded by dust rings).
157: \end{itemize}
158: Though plausible, these principles have not been proved rigorously so
159: far\footnote{In this context we refer to an new approach by
160: Masood-ul-Alam (2007).}. For proofs under special assumptions see
161: %\cite{Beig, Lindblom}.
162: Beig \& Simon (1992) and Lindblom \& Masood-ul-Alam (1994).
163:
164: \begin{figure}\centering
165: \includegraphics[scale=0.3]{f01.eps} %frueher: model.eps
166: \caption{Model: collisions of spherically symmetric stars or rigidly
167: rotating disks of dust}
168: \label{fig_model}
169: \end{figure}
170:
171: As illustrated in Fig.~\ref{fig_model} we will confine ourselves to two problems:
172: \begin{itemize}
173: \item[a)] head-on collisions of two identical spheres (stars) merging
174: into a single fluid ball and
175: \item[b)] head-on collisions of two identical disks of dust
176: (galaxies) merging into a single disk.
177: \end{itemize}
178:
179: We will be able to formulate \emph{necessary} conditions for the
180: formation of these balls or disks. Obviously, the conditions will
181: restrict the parameters of the initial configuration; a violation of the
182: conditions would necessarily lead to other final states such as to
183: black holes or central disks surrounded by rings. To express the
184: parameters of the admissible initial parameters --- the first goal of
185: this paper --- we have to solve the Einstein equations (numerically but)
186: rigorously and to make use of the conservation laws for baryonic mass
187: and angular momentum. There is no obstacle to an extension of the
188: method. One could start a systematic investigation of other possible
189: final states after the collision (black holes, black holes with rings
190: etc.) making use of symmetries, conservation laws and the heuristic
191: principles 1) and 2). An important point of the procedure would be the
192: stability analysis of the end products. As for our investigation, there
193: is important evidence from Newtonian gravity that rigidly or
194: differentially rotating disks of dust are unstable. Nevertheless we can
195: expect that \lq\lq stabilizing\rq\rq effects (pressure due to internal
196: kinetic energy) do not falsify our other goal --- to estimate the
197: maximal contribution of gravitational
198: radiation to the total energy loss $\Delta
199: M$. In general, the total energy loss calculated via the
200: comparison of initial and final equilibrium configurations is only an
201: upper limit for the energy loss (efficiency) due to gravitational
202: emission. (It includes energy loss due to non-gravitative radiation or
203: mass ejection during the collision). We will present a disk collision
204: model which is exclusively damped by gravitational radiation. The
205: resulting differentially rotating disk will be compared with a rigidly
206: rotating disk of the same baryonic mass and angular momentum formed from
207: the same initial disks under the additional influence of dissipative
208: processes in the matter (see \ref{DR}).
209: Thus we can compare the efficiencies of
210: the two forms of dissipation.
211:
212: In Sec.~\ref{Stars} we discuss, as introductory examples,
213: the merger of
214: two Schwarzschild stars and the collision of
215: two (Fermi gas)
216: neutron stars.
217: Sec.~\ref{Disks} contains the main part of this paper which is
218: dedicated to the investigation of disk collisions.
219: These discussions are based on a novel solution of the Einstein
220: equations describing the final configuration.
221: Here we continue the analysis of
222: a previous paper \cite{Hennig},
223: in which we discussed the collisions of rigidly rotating disks of
224: dust with parallel (or antiparallel) angular momenta under the simplifying
225: assumption that the final disk be again a \emph{rigidly} rotating (or
226: rigidly counterrotating) disk of
227: dust.
228: This assumption can only be justified if friction processes between the
229: disk rings provide for a constant angular velocity throughout the disk.
230: This model seems to be somewhat artificial and unsuited to determining the
231: contribution of gravitational radiation to the total energy
232: loss. Interestingly, our present investigation will show that the
233: frictional contribution to the total energy loss for colliding rigidly
234: rotating disks is comparably small.
235: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
236: \section{Star collisions}\label{Stars}
237:
238:
239: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
240: \subsection{Introductory example: Schwarzschild stars}
241: In order to demonstrate the method, we study the collision
242: of two Schwarzschild
243: stars, i.e. spherically symmetric perfect fluid stars with a constant
244: mass density, $\mu=\mathrm{constant}$.
245: Though not very realistic, this model illustrates the main steps
246: of the method.
247:
248: The matter of a Schwarzschild star is described by the perfect fluid
249: energy-momentum tensor
250: \begin{equation}\label{2}
251: T^{ij}=(\mu+p)u^i u^j+p g^{ij}
252: \end{equation}
253: with the pressure
254: \begin{equation}\label{2a}
255: p(r)=\frac{\sqrt{1-\frac{8\pi\mu}{3}r^2}
256: -\sqrt{1-\frac{8\pi\mu}{3}r_0^2}}{3\sqrt{1-\frac{8\pi\mu}{3}r_0^2}
257: -\sqrt{1-\frac{8\pi\mu}{3}r^2}}\mu,
258: \end{equation}
259: where $u^i$, $r$ and $r_0$ are the four-velocity,
260: the radial coordinate and the coordinate radius of
261: the star, respectively.
262: The interior Schwarzschild metric can be
263: written as
264: \begin{equation}\label{1}
265: \dd s^2=
266: \frac{\dd r^2}{1-\frac{8\pi\mu}{3}r^2}
267: +r^2(\dd\vartheta^2+\sin^2\vartheta\,\dd\varphi^2)
268: -\left(\frac{3}{2}\sqrt{1-\frac{8\pi\mu}{3}r_0^2}
269: -\frac{1}{2}\sqrt{1-\frac{8\pi\mu}{3}r^2}\right)\dd t^2,
270: \end{equation}
271: and the
272: exterior Schwarzschild solution is
273: \begin{equation}\label{2aa}
274: \dd s^2=\frac{\dd
275: r^2}{1-\frac{2M}{r}}
276: +r^2\left(\dd\vartheta^2+\sin^2\vartheta \dd\varphi^2\right)
277: -\left(1-\frac{2M}{r}\right)\dd t^2.
278: \end{equation}
279: Note that we use the normalized units where $c=1$ for the speed of
280: light and $G=1$ for Newton's gravitational constant.
281:
282: The gravitational mass $M$,
283: \begin{equation}\label{3}
284: M=\frac{4\pi\mu}{3}r_0^3,
285: \end{equation}
286: follows from the matching condition at the star's surface and the
287: baryonic mass $M_0$ is given by
288: \begin{equation}\label{4}
289: M_0=\int\limits_{t=t_0}\mu u^t\sqrt{-g}\,\dd r\dd\vartheta\dd\varphi
290: =4\pi\mu\int\limits_0^{r_0}\frac{r^2\dd
291: r}{\sqrt{1-\frac{8\pi\mu}{3}r^2}}.
292: \end{equation}
293:
294: Now we apply these formulae to the head-on collision of two stars.
295: Restricting ourselves to collisions of two identical
296: Schwarzschild stars we assume, that the final star
297: be again a Schwarzschild star and have the same mass density
298: (e.g. nuclear matter density),
299: \begin{equation}\label{4a}
300: \tilde\mu=\mu,
301: \end{equation}
302: where from now on tildes denote quantities after the collision.
303:
304: The conservation of baryonic mass during the collision
305: process
306: %\footnote{In all our considerations in this paper
307: %we will exclude mass loss by
308: %matter ejection or by radiation from
309: %thermonuclear reactions during the collision process.}
310: \begin{equation}\label{5}
311: \tilde M_0=2M_0,
312: \end{equation}
313: allows one to calculate the parameters of the final star as a function of
314: the initial parameters. With \eqref{3}, \eqref{4} and \eqref{4a} the
315: conservation equation \eqref{5} can be written as
316: \begin{equation}\label{6}
317: \arcsin\left(\sqrt{\frac{2M}{r_0}}\frac{\tilde r_0}{r_0}\right)
318: -\sqrt{\frac{2M}{r_0}}\frac{\tilde r_0}{r_0}
319: \sqrt{1-\frac{2M}{r_0}\frac{\tilde r_0^2}{r_0^2}}
320: =
321: 2\left(\arcsin\sqrt{\frac{2M}{r_0}}
322: -\sqrt{\frac{2M}{r_0}}
323: \sqrt{1-\frac{2M}{r_0}}\right),
324: \end{equation}
325: i.e. the radius ratio $\tilde r_0/r_0$ is a function of the
326: initial mass-radius ratio
327: $2M/r_0$. Hence, we may express the efficiency $\eta$
328: of conversion of mass into gravitational radiation,
329: \begin{equation}\label{7}
330: \eta=1-\frac{\tilde M}{2M}=1-\frac{1}{2}\left(\frac{\tilde r_0}{r_0}\right)^3,
331: \end{equation}
332: and the mass-radius ratio of the final star,
333: \begin{equation}\label{8}
334: \frac{2\tilde M}{\tilde r_0}=\frac{2M}{r_0}\left(\frac{\tilde r_0}{r_0}\right)^2,
335: \end{equation}
336: in terms of $2M/r_0$.
337:
338: \begin{figure}\centering
339: \includegraphics[scale=0.96]{f02a.eps}
340: \includegraphics[scale=0.96]{f02b.eps}\\[2ex]
341: \includegraphics[scale=0.96]{f02c.eps}
342: \caption{Parameter relations for colliding Schwarzschild stars:
343: The final mass-radius ratio $2\tilde M/\tilde r_0$, the
344: radius ratio $\tilde r_0/\tilde r$ and the efficiency
345: $\eta$ are plotted as functions of the initial
346: mass-radius ratio $2M/r_0$.
347: Dashed parts of the curves mark regions inaccessible due to the Buchdahl
348: inequality $2\tilde M/\tilde r_0< 8/9$.}
349: \label{fig_Schwarzschild}
350: \end{figure}
351:
352:
353: The resulting parameter relations are plotted in
354: Fig.~\ref{fig_Schwarzschild}. For Schwarzschild stars the coordinate
355: radius is restricted by the Buchdahl condition, i.e.
356: \begin{equation}\label{8a}
357: r_0>\frac{9}{8}\times 2M,
358: \quad
359: \tilde r_0>\frac{9}{8}\times 2\tilde M.
360: \end{equation}
361: As a consequence, the first plot
362: shows, that \lq\lq relativistic\rq\rq\ initial stars with
363: $2M/r_0>0.6482\dots$ can never merge into a new Schwarzschild star with the
364: same mass density $\mu$. The \lq\lq physical\rq\rq\ parts of the
365: parameter relations are shown as solid curves while the forbidden parts
366: are dashed.
367: According to the third plot the efficiency $\eta$ cannot exceed a
368: maximal value of $\eta_\textrm{max}\approx 19.7\%$.
369:
370: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
371: \subsection{Neutron stars: Completely degenerate ideal Fermi gas}
372:
373: In order to extend the discussion of the previous section
374: to a more realistic star model, we replace the equation of state
375: $\mu=\textrm{constant}$ by the equation for a completely
376: degenerate ideal Fermi gas of neutrons.
377:
378: The (interior) line element of a spherically symmetric star can be
379: written as
380: \begin{equation}\label{9}
381: \dd s^2=
382: \ee^{2\lambda(r)}\dd r^2
383: +r^2(\dd\vartheta^2+\sin^2\vartheta\,\dd\varphi^2)
384: -\ee^{2\nu(r)}\dd t^2
385: \end{equation}
386: and the matter is again described by the perfect fluid energy-momentum tensor
387: \begin{equation}\label{10}
388: T^{ij}=(\mu+p)u^i u^j+p g^{ij}.
389: \end{equation}
390: With the definition of a new metric function $m(r)$ by
391: \begin{equation}\label{11}
392: \ee^{2\lambda(r)}=\frac{1}{1-\frac{2m(r)}{r}},
393: \end{equation}
394: the field equations can be written in the TOV form,
395: %\cite{Shapiro}
396: see e.g. Shapiro \& Teukolsky (1983),
397: \begin{equation}\label{12}
398: \frac{\dd m}{\dd r}=4\pi r^2\mu, \quad m(0)=0
399: \end{equation}
400: \begin{equation}\label{13}
401: \frac{\dd p}{\dd r}=-\frac{m}{r^2}\mu\left(1+\frac{p}{\mu}\right)
402: \left(1+\frac{4\pi r^3 p}{m}\right)\left(1-\frac{2m}{r}\right)^{-1},
403: \quad p(0)=p_\textrm{c}
404: \end{equation}
405: \begin{equation}\label{14}
406: \frac{\dd \nu}{\dd r} = -\frac{1}{\mu}\frac{\dd p}{\dd r}
407: \left(1+\frac{p}{\mu}\right)^{-1},
408: \end{equation}
409: where $p_\textrm{c}$ is the pressure in the center of the star.
410:
411: We will solve
412: these equations for the
413: completely degenerate ideal fermi gas of neutrons
414: with the equation of state
415: %\cite{Shapiro}
416: [see e.g. Shapiro \& Teukolsky (1983)]
417: \begin{equation}\label{15}
418: p=c_1f(x),\quad \rho=c_2 x^3, \quad \mu=\rho+c_1 g(x),
419: \end{equation}
420: where
421: \begin{equation}\label{16}
422: f(x)=x(2x^2-3)\sqrt{1+x^2}+3\ln(x+\sqrt{1+x^2}),
423: \end{equation}
424: \begin{equation}\label{17}
425: g(x)=8x^3(\sqrt{1+x^2}-1)-f(x),
426: \end{equation}
427: \begin{equation}\label{18}
428: c_1=\frac{\pi m_\textrm{n}^4}{3h^3}, \quad c_2=\frac{8\pi m_\textrm{n}^4}{3h^3}
429: \end{equation}
430: with the neutron mass $m_\textrm{n}=1.6749286\times 10^{27}\,\mathrm{kg}$
431: and Planck's constant $h=6.626076\times 10^{-34}\,\mathrm{Js}$.
432: By
433: solving the TOV equations \eqref{12} and \eqref{13} with the
434: equation of state \eqref{15} for a sequence of values of the central density
435: one can calculate the corresponding radii of the stars as the first zero
436: $r_0$ of $p(r)$, their gravitational mass from $M=m(r_0)$,
437: and their baryonic mass as
438: \begin{equation}\label{19}
439: M_0=4\pi\int\limits_0^{r_0}\frac{\rho(r)r^2\dd r}{\sqrt{1-\frac{2m(r)}{r}}}.
440: \end{equation}
441: The resulting mass-radius relations are shown in the first plot of
442: Fig.~\ref{fig_Neutronenstern}.
443:
444:
445: \begin{figure}\centering
446: \includegraphics[scale=0.98]{f03a.eps}
447: \includegraphics[scale=0.98]{f03b.eps}\\[2ex]
448: \includegraphics[scale=0.98]{f03c.eps}
449: \includegraphics[scale=0.98]{f03d.eps}
450: %frueher: Neutronenstern_d.eps
451: \caption{Parameter relations for the collisions of neutron stars made up
452: of degenerate neutrons [cf. \eqref{15}]. First
453: plot: mass-radius relations for the baryonic mass $M_0$ and the
454: gravitational mass $M$. Second plot: initial radius $r_0$ and final
455: radius $\tilde r_0$ as functions of the mass-radius ratio $2M/r_0$.
456: Third plot: change of the coordinate
457: radius. Fourth plot: efficiency $\eta$ compared to the efficiency of
458: the collision of Schwarzschild stars (dashed curve). }
459: \label{fig_Neutronenstern}
460: \end{figure}
461:
462:
463: Again the baryonic mass is an invariant of the collision, i.e.
464: \begin{equation}\label{20}
465: \tilde M_0=2M_0
466: \end{equation}
467: for the collision of two identical initial stars.
468: This equation has to be analysed together with the mass-radius relations.
469: (Thereby we
470: take into account only stars in the monotonic decreasing part of the
471: mass-radius relation $M_0(r_0)$.) The resulting
472: parameter relations are shown in
473: the remaining plots of Fig.~\ref{fig_Neutronenstern}.
474: For the maximum of the efficiency one finds
475: $\eta_\textrm{max}\approx 2.3\%$, i.e. a
476: comparably small value in view of the maximal efficiency
477: $\eta_\textrm{max}\approx 19.7\%$
478: for collisions of
479: Schwarzschild stars. The reason is the relatively small maximal mass of
480: $M_\textrm{max}\approx 0.7 M_\odot$
481: permitted by the equation of state \eqref{15} that excludes
482: highly relativistic values for the mass-radius ratio
483: $2M/r_0$. However, compared to Schwarzschild stars with the same parameter
484: $2M/r_0$ the collisions of neutron stars are more efficient,
485: cf. the last plot in
486: Fig.~\ref{fig_Neutronenstern}.
487:
488: Another difference is the change of the coordinate radii.
489: While two Schwarzschild stars merge into a Schwarzschild star with a
490: coordinate radius bigger than the initial radius,
491: $\tilde r_0/r_0>1$ (cf. Fig.~\ref{fig_Schwarzschild}),
492: the resulting neutron star is smaller than the initial neutron stars,
493: $\tilde r_0/r_0<1$ (cf. Fig.~\ref{fig_Neutronenstern}).
494:
495:
496: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
497: \section{Disk collisions}\label{Disks}
498:
499: Collisions of disks of dust require more effort. In particular,
500: the discussion of the final equilibrium state is based on a solution
501: of a free boundary value problem to the Einstein equations.
502: At the first glance, the conservation laws for baryonic mass and angular
503: momentum are not sufficient to formulate a complete set of boundary
504: conditions for the configuration after the collision. However, excluding
505: non-gravitational dissipation, we may
506: replace the global conservation laws, as used in Sec.~\ref{Stars}, by
507: local ones.
508: Due to the geodesic motion of dust particles, the baryonic mass and the
509: angular momentum of each of the rings forming the disk are conserved
510: separately, see Fig.~\ref{fig_local_conservation}. Using such \emph{local}
511: conservation laws we will be able to solve
512: (numerically) the boundary value problem for the final state after the
513: head-on collision of two aligned rigidly rotating disks
514: of dust with parallel angular momenta, cf. Fig.~\ref{fig_model}.
515:
516: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
517:
518: \subsection{Initial disks: Rigidly rotating disks of dust}
519:
520: The free boundary value problem for the relativistic rigidly rotating
521: disk of dust (RR disk) was
522: discussed by Bardeen and Wagoner (1969, 1971) using approximation methods
523: and
524: analytically solved in terms of ultraelliptic theta functions
525: by Neugebauer and Meinel (1993, 1994, 1995)
526: using the Inverse Scattering
527: Method.
528: The line element of the stationary (Killing vector: $\xi^i$) and
529: axisymmetric (Killing vector: $\eta^i$) space-time may be written
530: in the Weyl-Lewis-Papapetrou standard form
531: \begin{equation}\label{21}
532: \dd s^2=\ee^{-2U}[\ee^{2k}(\dd
533: \rho^2+\dd\zeta^2)+\rho^2\dd\varphi^2]-\ee^{2U}(\dd t+a\dd\varphi)^2,
534: \quad \xi^i=\delta^i_t,\quad \eta^i=\delta^i_\varphi,
535: \end{equation}
536: where the metric potentials
537: $U=U(\rho,\zeta)$, $k=k(\rho,\zeta)$ and $a=a(\rho,\zeta)$ are
538: given in terms of ultraelliptic theta functions.
539:
540: The matter of the disk of dust is described by the energy-momentum tensor
541: \begin{equation}\label{22}
542: T^{ij}=\varepsilon(\rho)\delta(\zeta)u^i u^j,
543: \end{equation}
544: where $\varepsilon(\rho)\delta(\zeta)$ is the mass density with
545: $\delta(\zeta)$ as Dirac's $\delta$-distribution. Due to the
546: symmetries, the
547: four-velocity of the dust particles is a linear combination of the two
548: killing vectors,
549: \begin{equation}\label{22a}
550: u^i=\ee^{-V_0}(\xi^i+\Omega_0\eta^i),\quad u^i u_i=-1,
551: \end{equation}
552: whence
553: \begin{equation}\label{22b}
554: (\xi^i+\Omega_0\eta^i)(\xi_i+\Omega_0\eta_i)=-\ee^{2V_0},
555: \end{equation}
556: where $\Omega_0$ is the angular velocity of the particles forming the
557: disk and $V_0$ is a redshift parameter. Rigid rotation means
558: $\Omega_0=\textrm{constant}$ in the disk.
559: Since dust particles move geodesically
560: this assumption implies $V_0=\textrm{constant}$ in the disk. Hence, the boundary
561: condition \eqref{22b} and as a consequence the RR disk solution contains
562: two constant parameters.
563: Alternatively to $\Omega_0$ and $V_0$, one may choose the
564: coordinate radius $\rho_0$ of the disk and a centrifugal parameter
565: $\mu=2\Omega_0^2\rho_0^2\ee^{-2V_0}$ [$\mu\to 0$ turns out to be the
566: Newtonian limit and $\mu\to 4.62966\dots$ the ultrarelativistic limit,
567: where the disk approaches the extreme Kerr black hole,
568: cf.
569: %\cite{Neugebauer2} and \cite{Neugebauer4}
570: Neugebauer \& Meinel (1994) and Neugebauer et al. (1996)
571: for these
572: and further properties].
573: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
574: \subsection{Final disk: Differentially rotating disk of dust}
575: \label{DR}
576:
577: In a previous paper \cite{Hennig} we discussed head-on collisions
578: of two (identical) rigidly rotating disks of dust merging into one
579: \emph{rigidly} rotating disk of dust. The model excluded mass ejection
580: and made use of the conservation of baryonic mass and angular momentum
581: (axisymmetry). From a thermodynamic point of view rigid rotation of the
582: final disk means thermodynamic equilibrium, which is a result of
583: dissipative processes during the dynamical phase. Hence, the energy
584: difference between the initial state (two separated disks) and the final
585: state (one rigidly rotating disk) is influenced by irreversible
586: processes in the matter and outgoing electromagnetic radiation as well
587: as by emission of gravitational waves. The intention of this paper is to
588: compare the contribution of these two effects by calculating the end
589: product of a purely gravitational collision process which we expect to
590: be a \emph{differentially} rotating disk of dust. Note that our
591: thermodynamic analysis enables us to formulate \emph{necessary}
592: conditions for the parameters of the initial disks ($\mu$ restricted) to
593: permit the formation of a final \emph{disk}. To obtain \emph{sufficient}
594: conditions one would have to solve the Einstein equations for the
595: time-dependent transition phase, which is outside the scope of this paper.
596:
597: In the next subsection we will give a brief summary of the previous
598: paper.
599: After that, we will see that the \emph{local} conservation of baryonic mass
600: and angular
601: momentum is sufficient to calculate the final
602: differentially rotating disk (numerically).
603: Differentially rotating disks with arbitrary rotation law
604: have already been studied \cite{Ansorg1, Ansorg2}.
605: The point made here is that we are able to formulate a physically
606: motivated rotation law as a result of a collision process.
607:
608:
609:
610: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
611: \subsubsection{Formation of rigidly rotating disks}
612: \label{AssumptionRR}
613:
614: For the formation of an RR disk from two colliding RR disks under the
615: influence of friction processes
616: the conservation equations for baryonic mass and angular momentum,
617: \begin{equation}\label{23}
618: \tilde M_0=2M_0,\quad \tilde J=2J,
619: \end{equation}
620: are sufficient to calculate the parameters of the final disks as functions
621: of the initial parameters. These equations and
622: explicit formulae connecting the gravitational mass $M$,
623: the baryonic mass $M_0$ and the angular momentum $J$ of the RR disk
624: allowed us to calculate
625: the efficiency $\eta^{\textrm{RR}}=1-\tilde M/2M$
626: as a function of the initial centrifugal
627: parameter $\mu$, cf. Fig.~\ref{fig_efficiency(RR)}. It should be
628: emphasized once again,
629: that this efficiency measures the total energy loss including
630: friction. Therefore $\eta$ is only an \emph{upper limit} for the energy of the
631: gravitational emission.
632: We obtained a maximal value of $\eta^{\textrm{RR}}_{\textrm{max}}\approx
633: 23.8\%$
634: \cite{Hennig}.
635:
636:
637: \begin{figure}\centering
638: \includegraphics[scale=1]{f04.eps}
639: \caption{The efficiency $\eta_{\textrm{RR}}$
640: for the formation of an RR disk from two initial
641: RR disks as a function of the centrifugal parameter $\mu$ of the initial
642: disks. $\eta^{\textrm{RR}}$ is an upper limit for the energy
643: loss due to gravitational
644: radiation.}
645: \label{fig_efficiency(RR)}
646: \end{figure}
647:
648: Furthermore, it turned out that
649: the formation of RR disks from two colliding RR disks is only possible
650: for a rather restricted interval
651: $0<\mu<1.954\dots$ of the initial centrifugal parameter $\mu$.
652: If $\mu$
653: exceeds this limit, the
654: collision must lead to other final states,
655: e.g. black holes or black holes surrounded by matter rings.
656:
657: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
658: \subsubsection{Local conservation equations}
659:
660: \begin{figure}\centering
661: \includegraphics[scale=0.6]{f05.eps}
662: %frueher: local_conservation.eps
663: \caption{Illustration of the local conservation equations.
664: Two corresponding rings of the initial
665: disks merge into a ring in the final disk. The
666: baryonic mass $\dd M_0$ and the angular momentum $\dd J$ of
667: these rings are conserved.}
668: \label{fig_local_conservation}
669: \end{figure}
670:
671: We now turn to the main goal of this paper and analyse the formation of a
672: disk of dust under the influence of gravitational forces as the only
673: form of interaction.
674: Comparing the resulting \emph{differentially} rotating disk of dust with
675: the \emph{rigidly} rotating disk of the same baryonic mass $M_0$ and
676: angular momentum $J$ formed from the same initial disks
677: we may separate gravitational damping due to the emission of
678: gravitational waves from frictional processes in the matter.
679:
680:
681: We may interpret a disk of dust as a superposition of infinitesimally
682: thin dust rings.
683: Considering the geodesic motion of a single mass element,
684: one can show that for corresponding rings in the two initial disks
685: (see Fig.~\ref{fig_local_conservation})
686: the baryonic mass and the angular
687: momentum are conserved,
688: \begin{equation}\label{24}
689: \dd\tilde M_0=2\dd M,\quad \dd\tilde J=2\dd J,
690: \end{equation}
691: i.e. the baryonic masses $\dd M_0$ and the angular momenta
692: $\dd J$ of the rings with radius $\rho$, taking up the interval
693: [$\rho,\rho+\dd\rho]$,
694: in each of the two initial disks sum up to
695: $\dd\tilde M_0=2\dd M_0$ and $\dd\tilde J=2\dd J$
696: of the corresponding ring in the final disk (with radius $\tilde\rho$,
697: taking up the interval $[\tilde\rho,\tilde\rho+\dd\tilde\rho]$),
698: cf. Fig.~\ref{fig_local_conservation}.
699: It should be emphasized that the local conservation laws \eqref{24}
700: would be violated by dissipative processes in the matter or,
701: mathematically speaking, by dissipative terms in the total
702: energy-momentum tensor as the source of the Einstein equations during
703: the collision phase.
704: Having reached a final equilibrium configuration (e.g. a \emph{rigidly}
705: rotating disk of dust) the system \lq\lq forgets\rq\rq\ the dissipative
706: terms and behaves like cold dust with an energy momentum tensor of the
707: form \eqref{22}. During the interaction phase, angular momentum will be
708: transported within the disk by viscous forces and only the \emph{total}
709: angular momentum (axisymmetry!) and the total baryonic mass are
710: conserved \eqref{23}. The \emph{ring-wise} conservation of baryonic mass
711: and angular momentum \eqref{24} is characteristic for purely
712: gravitational damping processes. They arise from collision processes
713: governed by an energy-momentum tensor of dust without dissipative
714: terms. In this case the geodesic motion of the volume elements
715: implies the conservation of baryonic mass and angular momentum in each
716: volume element and therefore implies \eqref{24}.
717:
718: Eq.~\eqref{24} provides us with a subset of the boundary conditions
719: to be discussed in the next subsection.
720: It will turn out that these conditions, together with conditions
721: resulting from the field equations, determine a unique solution of the
722: Einstein equations describing a final disk with differential rotation
723: (DR disk) as the end product of the collision process.
724:
725:
726: \subsubsection{Boundary value problem for the final DR disk}
727: \label{BVP}
728:
729: The line element \eqref{21}, which may also be used to describe
730: any axisymmetric and stationary \emph{differentially}
731: rotating disk, can be reformulated to give
732: \begin{equation}\label{25}
733: \dd \tilde
734: s^2=\ee^{2\tilde\kappa}(\dd\tilde\rho^2+\dd\tilde\zeta^2)+\tilde\rho^2\ee^{-2\tilde\nu}(\dd\tilde\varphi-\tilde\omega\dd\tilde
735: t)^2-\ee^{2\tilde\nu}\dd \tilde t^2,
736: \end{equation}
737: where the usage of the functions $\tilde\kappa$, $\tilde\nu$ and $\tilde\omega$
738: [instead of $\tilde U$, $\tilde k$ and $\tilde a$ as in \eqref{21}]
739: avoids numerical issues
740: with ergospheres (where $\ee^{2\tilde U}<0$). According to \eqref{22}
741: the energy-momentum
742: tensor is
743: \begin{equation}\label{25a}
744: \tilde T^{ij}=\tilde\varepsilon(\tilde\rho)\delta(\tilde\zeta)\tilde
745: u^i\tilde u^j
746: \end{equation}
747: and the four-velocity is again
748: [cf. \eqref{22a}] a linear combination of the killing
749: vectors,
750: \begin{equation}\label{25b}
751: \tilde u^i=\ee^{-\tilde V}(\tilde\xi^i+\tilde\Omega\tilde\eta^i),
752: \end{equation}
753: where $\tilde V=\tilde V(\tilde\rho)$ and
754: $\tilde\Omega=\tilde\Omega(\tilde\rho)$ are functions of $\tilde\rho$
755: [constancy of $\tilde V$ and $\tilde\Omega$ defines rigid rotation,
756: cf.~\eqref{22a}].\footnote{All
757: quantities of the final DR disk are tilded.}
758:
759: The vacuum field equations for $\tilde\nu$ and
760: $\tilde\omega$ are
761: %(cf. \cite{BlackHoles})
762: [cf. Bardeen (1973)]
763: \begin{equation}\label{26}
764: \bigtriangleup_1\tilde\nu
765: =\frac{\tilde\rho^2}{2}\ee^{-4\tilde\nu}
766: (\tilde\omega_{,\tilde\rho}^{\ 2}+\tilde\omega_{,\tilde\zeta}^{\ 2}),
767: \quad
768: \bigtriangleup_3\tilde\omega
769: =4(\tilde\nu_{,\tilde\rho}\tilde\omega_{,\tilde\rho}
770: +\tilde\nu_{,\tilde\zeta}\tilde\omega_{,\tilde\zeta}),
771: \end{equation}
772: with
773: \begin{equation}\label{27}
774: \bigtriangleup_n:=\partial_{\tilde\rho}^2+
775: \partial_{\tilde\zeta}^2+\frac{n}{\tilde\rho}\partial_{\tilde\rho}.
776: \end{equation}
777: The matter appears only in the boundary conditions along the disk
778: ($\tilde\zeta=0$, $\tilde\rho<\tilde\rho_0$),
779: \begin{equation}\label{28}
780: \left.\tilde\nu_{,\tilde\zeta}\right|_{\tilde\zeta=0^+}
781: =2\pi\tilde\sigma\frac{1+\tilde v^2}{1-\tilde v^2},
782: \end{equation}
783: \begin{equation}\label{28a}
784: \left.\tilde\omega_{,\tilde\zeta}\right|_{\tilde\zeta=0^+}
785: =-8\pi\tilde\sigma\frac{\tilde\Omega-\tilde\omega}{1-\tilde v^2},
786: \end{equation}
787: \begin{equation}\label{29}
788: \tilde\rho(\tilde\Omega-\tilde\omega)^2=(1+\tilde
789: v^2)\ee^{4\tilde\nu}\tilde\nu_{,\tilde\rho}
790: +\tilde\rho^2(\tilde\Omega-\tilde\omega)\tilde\omega_{,\tilde\rho},
791: \end{equation}
792: where
793: \begin{equation}\label{30}
794: \tilde v:=\tilde\rho\ee^{-2\tilde\nu}(\tilde\Omega-\tilde\omega),\quad
795: \tilde\sigma:=\tilde\varepsilon\ee^{2\tilde\kappa}.
796: \end{equation}
797: Thus we have to deal with a boundary value problem for Einstein's vacuum
798: equations.
799:
800: As already mentioned, the local conservation equations \eqref{24}
801: of the previous subsection lead to
802: additional boundary conditions along the disk.
803: From $\dd\tilde M_0=2\dd M_0$ with $\dd M_0=2\pi\sigma\ee^{-V_0}\rho\dd\rho$
804: and $\dd\tilde M_0=2\pi\tilde\sigma\ee^{-\tilde V}\tilde\rho\dd\tilde\rho$
805: we obtain
806: \begin{equation}\label{31}
807: \tilde\sigma=2\sigma\frac{\rho\ee^{V_0}}{\tilde\rho\ee^{\tilde
808: V}}\frac{\dd\rho}{\dd \tilde\rho}.
809: \end{equation}
810: Likewise, $\dd\tilde J=2\dd J$ with $\dd J=2\pi\sigma\ee^{-V_0}u^i\eta_i\rho\dd\rho$
811: and
812: $\dd\tilde J
813: =2\pi\tilde\sigma\ee^{-\tilde V}\tilde
814: u^i\tilde\eta_i\tilde\rho\dd\tilde\rho$,
815: $u^i\eta_i=\rho v\ee^{-V_0}$ and
816: $\tilde u^i\tilde\eta_i=\tilde\rho\tilde v\ee^{-\tilde V}$
817: leads to
818: \begin{equation}\label{32a}
819: \tilde\rho\tilde v\ee^{-\tilde V}=\rho v\ee^{-V_0}.
820: \end{equation}
821: The function $\tilde V(\tilde\rho)$
822: can be calculated from $\tilde u^i\tilde u_i=-1$,
823: \begin{equation}\label{33}
824: \ee^{2\tilde V}=(1-\tilde v^2)\ee^{2\tilde\nu}.
825: \end{equation}
826: The remaining boundary conditions describe the behaviour at infinity, where the
827: metric approaches the flat Minkowski metric,
828: \begin{equation}\label{33a}
829: \tilde\kappa=\tilde\nu=\tilde\omega=0,
830: \end{equation}
831: and in the plane $\tilde\zeta=0$ outside the disk
832: ($\tilde\rho>\tilde\rho_0$), where \eqref{28} and
833: \eqref{28a} lead to vanishing normal derivatives,
834: \begin{equation}\label{33b}
835: \left.\tilde\nu_{,\tilde\zeta}\right|_{\tilde\zeta=0^+}=0,\quad
836: \left.\tilde\omega_{,\tilde\zeta}\right|_{\tilde\zeta=0^+}=0.
837: \end{equation}
838: In addition we have to ensure
839: regularity along the axis of symmetry $\tilde\rho=0$.
840:
841: Eqs.~\eqref{26}, \eqref{28}-\eqref{29} and \eqref{31}, \eqref{32a} form
842: a complete set of equations to determine the unknown
843: functions uniquely:
844: There are two two-dimensional functions,
845: $\tilde\nu(\tilde\rho,\tilde\zeta)$ and
846: $\tilde\omega(\tilde\rho,\tilde\zeta)$, which have to satisfy the
847: two elliptic partial differential
848: equations \eqref{26} with the boundary conditions
849: \eqref{28} and \eqref{28a}, and three additional one-dimensional
850: functions in the disk, $\tilde\Omega(\tilde\rho)$,
851: $\tilde\sigma(\tilde\rho)$, $\rho(\tilde\rho)$, which have to obey
852: the three boundary conditions \eqref{29}, \eqref{31} and \eqref{32a}.
853: (The metric function
854: $\tilde\kappa$ can be calculated by a line integral afterwards, but is
855: not needed for the computation of the efficiency $\eta$ in our
856: collision scenario.)
857:
858: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
859: \subsubsection{Numerical method}
860:
861: \begin{figure}\centering
862: \includegraphics[scale=1]{f06.eps}
863: \caption{The coordinate transformation \eqref{34}, \eqref{35} maps the
864: part $\tilde\zeta\ge 0$ of the $\tilde\rho$-$\tilde\zeta$-plane to
865: a unit square in the $s$-$t$-plane. $\mathcal E$ denotes
866: the equatorial plane outside the disk, $\tilde\zeta=0$,
867: $\tilde\rho>\tilde\rho_0$.}
868: \label{fig_Koordinatentransformation}
869: \end{figure}
870:
871: In order to prepare numerical investigations
872: we will map the region $0\le \tilde\rho \le \infty$, $0\le \tilde\zeta\le
873: \infty$ to a unit square thus reaching a compactification of infinity,
874: cf. Fig.~\ref{fig_Koordinatentransformation}.
875: (Due to the reflection symmetry with respect
876: to the plane
877: $\tilde\zeta=0$ we can restrict ourselves to the region $\tilde\zeta\ge 0$.)
878: To do this we introduce in a first step elliptical coordinates
879: \begin{equation}\label{34}
880: \tilde \rho=\sqrt{(1+\xi^2)(1-\eta^2)},\quad \tilde\zeta=\xi\eta,\quad
881: \xi\in[0,\infty],\quad \eta\in[0,1]
882: \end{equation}
883: (without loss of generality,
884: we may choose units where $\tilde\rho_0=1$).
885: In a second step we stretch the coordinates by the transformation
886: \begin{equation}\label{35}
887: \xi=\cot\left(\frac{\pi}{2}s\right),\quad \eta=\sqrt{1-t},\quad
888: s\in[0,1], \quad t\in[0,1].
889: \end{equation}
890: The coordinates $s$ and $t$ form a unit square with the
891: following boundaries,
892: \begin{equation}
893: \begin{array}{ll}
894: s=0: & \infty\\
895: s=1: & \textrm{disk},\ \tilde\rho\le 1,\ \tilde\zeta=0\\
896: t=0: & \textrm{axis of symmetry},\ \tilde\rho=0\\
897: t=1: & \textrm{disk plane $\mathcal E$
898: outside the matter},\ \tilde\rho>1,\ \tilde\zeta=0.
899: \end{array}
900: \nonumber
901: \end{equation}
902: The unknown functions in the boundary value problem are analytic
903: functions in this square (as is known for the case of Maclaurin disks or
904: the RR disks). Hence, it is convenient to use spectral methods for the
905: numerical solution of the boundary value problem.
906: We expand the unknown potentials in terms of
907: Chebyshev polynomials $T_j$ to a predetermined
908: order in the form
909: \begin{equation}\label{36}
910: f(s,t)=\sum\limits_{j,k}c_{jk} T_j(2s-1) T_k(2t-1)\quad \textrm{or}\quad
911: f(t)=\sum\limits_{k}c_k T_k(2t-1)\quad\textrm{(boundary)}
912: \end{equation}
913: and formulate the Einstein equations at the extrema of the Chebyshev
914: polynomials. This
915: leads to an algebraic system of equations for the Chebyshev coefficients (or,
916: alternatively, for the values of the potentials at these points) that
917: can be solved with the Newton-Raphson method. The iteration starts with an
918: initial \lq\lq guessed\rq\rq\ solution (for example the Newtonian
919: approximation, see Sec.~\ref{Newtonian} below).
920:
921: The calculations show a decreasing accuracy of the numerical solution
922: for increasingly large values of the initial parameter $\mu$.
923: The reason are large
924: gradients of the metric potentials for strong relativistic DR disk which
925: make the Chebyshev approximation more costly.
926: To reach a better convergence we perform an additional coordinate
927: transformation
928: \begin{equation}\label{36a}
929: s=\frac{\sinh(\delta\cdot\tilde s)}{\sinh(\delta)}
930: \end{equation}
931: introducing a new coordinate $\tilde s$, where $\delta$ is a suitably
932: chosen parameter. As shown in
933: %\cite{Rings}
934: Ansorg~\&~Petroff (2005), this transformation smooths the
935: gradients of the metric functions.
936: The convergence is illustrated in Fig.~\ref{Konvergenztest}.
937:
938: \begin{figure}\centering
939: \includegraphics[scale=1.1]{f07.eps}
940: \caption{Convergence properties of the numerical code for the example
941: $\eta^{\textrm{DR}}$. The values of the efficiency $\eta^{\textrm{DR}}$ for
942: different orders $n_s=n_t=n$ of the Chebyshev expansion
943: are related to the order $n=32$. The plot shows
944: $\left| 1-{\eta^{\textrm{DR}}_n}/{\eta^{\textrm{DR}}_{32}}\right|$ as
945: function of $n$.}
946: \label{Konvergenztest}
947: \end{figure}
948:
949: %%%%%%%%%%%%%%%%%%%%%%%%%%%
950: \subsubsection{Results}\label{results}
951:
952: \begin{figure}\centering
953: \includegraphics[scale=0.3]{f08.eps} %frueher: model_Scheiben.eps
954: \caption{Two models for disk collisions:\newline
955: (A) Under the influence of
956: a small amount of friction, RR disks merge again into an RR disk. This
957: scenario was discussed in
958: %\cite{Hennig}
959: Hennig \& Neugebauer (2006), see Sec.~\ref{AssumptionRR}.
960: \newline
961: (B) In the absence of friction, the same RR disks merge into
962: a DR disk.
963: Allowing for friction afterwards, the system would again arrive at the RR
964: disk of scenario (A) after a sufficiently long time.
965: %\newline
966: %The baryonic mass and the angular momentum are conserved locally in
967: %scenario (B). However, in scenario (A), angular momentum will be
968: %transported within the disc by viscous forces.
969: }
970: \label{model_Scheiben}
971: \end{figure}
972:
973:
974: Using this numerical algorithm, we are able to solve the
975: boundary value problem for the final DR disk.
976: In particular, we could calculate, for each value of the initial
977: parameter $\mu$,
978: all \emph{metric coefficients} of this final disk.
979: However, we will restrict ourselves to the discussion
980: of the relations between the initial and final \emph{parameters} and the
981: \emph{efficiency} of the collision process.
982: Especially, we will compare the
983: final DR disk with an RR disk having the same baryonic mass and
984: angular momentum.
985: The point made here is that such a rigidly rotating disk represents the
986: state of \lq\lq thermodynamic equilibrium\rq\rq\ for disks of dust as
987: the end point of their thermodynamic evolution.
988: As sketched in Fig.~\ref{model_Scheiben},
989: there are at least two possibilities for the formation
990: of this final RR disk:
991: The direct process (A) including friction from the beginning
992: or the equivalent thermodynamic process (B) where, in a first step, a
993: \emph{differentially rotating} disk is formed (by gravitational damping
994: alone, no friction)
995: and, in a second step, the angular velocity becomes
996: \emph{constant} (due to friction).
997: Note that baryonic mass and angular momentum are conserved in both
998: processes.
999: By comparing (A) and (B) we may
1000: extract the contribution of friction in scenario~(A).
1001:
1002:
1003: In the following discussion, tilded quantities, as before, belong to the final
1004: DR disk of scenario (B) in Fig.~\ref{model_Scheiben},
1005: a superscript \lq\lq RR\rq\rq\ denotes
1006: quantities of the final RR disk in scenario (A) and the centrifugal
1007: parameter $\mu$ without any additions characterizes the initial RR disks.
1008:
1009: \begin{figure}\centering
1010: \includegraphics[scale=0.96]{f09a.eps}
1011: \includegraphics[scale=0.96]{f09b.eps}\\[2ex]
1012: \includegraphics[scale=0.96]{f09c.eps}
1013: \includegraphics[scale=0.96]{f09d.eps}\\[2ex]
1014: \includegraphics[scale=0.96]{f09e.eps}
1015: \includegraphics[scale=0.96]{f09f.eps}
1016: \caption{Parameter relations for the collision of RR disks. We performed
1017: numerical calculations for values of the initial centrifugal parameter
1018: $\mu$ in the invervall $[0,1.9]$. The dotted part of the curve in the
1019: last plot is an extrapolation for larger $\mu$. This extrapolation and
1020: the rapidly growing redshift in the second last plot indicate, that the
1021: initial parameter $\mu$ in scenario (B)
1022: is limited (approximately, or perhaps even
1023: exactly) to the same interval as in scenario (A), $0<\mu<1.954\dots$,
1024: see Sec.~\ref{AssumptionRR}.}
1025: \label{fig_DR}
1026: \end{figure}
1027:
1028:
1029: The rotation curve of the final DR disk, i.e. its
1030: (normalized) angular velocity $\tilde\Omega\tilde\rho_0$
1031: as a function of the (normalized) radius $\tilde\rho/\tilde\rho_0$
1032: is shown in the first two plots of
1033: Fig.~\ref{fig_DR}. For small parameters $\mu$ (post-Newtonian regime)
1034: the function
1035: $\tilde\Omega\tilde\rho_0$ is almost constant (first plot).
1036: Interestingly, strongly relativistic disks ($\mu\gtrsim 1.5$)
1037: show the same property
1038: (second plot). Moreover,
1039: $\tilde\Omega\tilde\rho_0$ tends to zero in the
1040: ultrarelativistic limit in analogy to
1041: the relation $\Omega^{\textrm{RR}}\rho^{\textrm{RR}}_0\to 0$
1042: which holds for RR disks in the
1043: ultrarelativistic limit $\mu^{\textrm{RR}}\to 4.62966\dots$
1044:
1045: The \lq\lq centrifugal parameter\rq\rq\
1046: $\tilde\mu=2\tilde\Omega^2\tilde\rho_0^2\ee^{-\tilde V}=\mu(\tilde\rho)$ is
1047: shown in the third plot of Fig.~\ref{fig_DR}. Like the angular
1048: velocity, $\tilde\mu$ is almost constant for small $\mu$. For
1049: strongly relativistic DR disks, $\tilde\mu$ in the center of the disk
1050: exceeds the limit $\mu^{\textrm{RR}}_{\textrm{max}}=4.62966\dots$ of RR disks.
1051:
1052:
1053: The fourth plot of Fig.~\ref{fig_DR} shows the quantity
1054: $\tilde\Omega\tilde M$ as a function of $\rho/\rho_0$.
1055: For strongly relativistic DR disks
1056: $\tilde\Omega\tilde M$ becomes constant and approaches the limit $0.5$.
1057: On the other hand, this is a characteristic value for extreme Kerr
1058: black holes where
1059: $\Omega_\textrm{H}M_\textrm{BH}=0.5$ ($\Omega_\textrm{H}$: angular
1060: velocity of the horizon, $M_\textrm{BH}$: black hole mass). Indeed, one
1061: can show that there is a phase transition between RR disks and Kerr
1062: black holes \cite{Bardeen2, Neugebauer2}.
1063: This inspires the conjecture that the DR disk exhibits the same
1064: phase transition.
1065: There is no obstacle for a (numerical) proof of this
1066: assumption in principle.
1067: To extend our present code to study the parametric collapse
1068: of the DR disk including the formation of a horizon we would have
1069: to follow the
1070: ideas of Bardeen and Wagoner (1971)
1071: %\cite{Bardeen2}
1072: who analysed this problem
1073: for the RR disks. However, such investigations are outside the scope of
1074: this paper.
1075:
1076:
1077: The fifth plot of Fig.~\ref{fig_DR} shows the redshift $\tilde z$ for a
1078: photon emitted from the disk center
1079: as a function of the initial centrifugal parameter $\mu$.
1080: For
1081: increasing values of $\mu$ (relativistic DR disks) $\tilde z$ grows rapidly.
1082:
1083:
1084:
1085: An important result is the efficiency $\eta^{\textrm{DR}}$ of the formation
1086: of DR disks which measures the amount of energy converted into
1087: gravitational radiation.
1088: The difference $\eta^{\textrm{F}}=\eta^{\textrm{RR}}-\eta^{\textrm{DR}}$ as
1089: shown in the last plot of Fig.~\ref{fig_DR} compares this value with
1090: the efficiency $\eta^{\textrm{RR}}$ of the RR disk forming process as
1091: sketched in scenario (A) of Fig.~\ref{model_Scheiben}.
1092: Thereby, $\eta^{\textrm{F}}$ is the part of energy lost due to
1093: friction during the formation of a final RR disk.
1094: We find
1095: $\eta^{\textrm{F}}=\eta^{\textrm{RR}}-\eta^{\textrm{DR}}<1.5\times 10^{-4}$,
1096: i.e. the contribution of friction is extremely small,
1097: $ \eta^{\textrm{F}}\ll\eta^{\textrm{RR}}$,
1098: such that the gravitational radiation dominates the collision process (A).
1099:
1100:
1101: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1102: \subsubsection{Analytical treatment of the Newtonian limit}
1103: \label{Newtonian}
1104:
1105: Our numerical investigations have shown that the angular velocity of
1106: the final DR disk
1107: becomes closer and closer to a constant over the whole range of
1108: $\tilde\rho/\tilde\rho_0$ as the centrifugal parameter $\mu$ tends to
1109: zero (cf. the first plot of Fig.~\ref{fig_DR}).
1110: This leads one to suspect that a final disk with a
1111: \emph{strictly} constant
1112: angular velocity will solve the boundary value problem
1113: as discussed in Sec.~\ref{BVP} in \emph{Newtonian}
1114: theory. Interestingly, we can treat this problem analytically. This will
1115: now be demonstrated. Strictly speaking, there is no gravitational
1116: radiation in Newton's theory. However, this Newtonian
1117: boundary value problem can be seen as the limit of a sequence of
1118: relativistic collisions with decreasing $\mu$, all reaching a final
1119: equilibrium state due to gravitational emission.
1120: Moreover, the Newtonian solution can be used as a
1121: starting point for the iterative calculation of the final relativistic
1122: DR disk.
1123:
1124:
1125: Since the Newtonian limit of the RR disk is the Maclaurin
1126: disk we have to study the collision of two identical Maclaurin
1127: disks using the local conservation laws~\eqref{24}.
1128: The Newtonian potential $\tilde U$ of the final disk is a
1129: solution of the Poisson equation
1130: \begin{equation}\label{N1}
1131: \bigtriangleup\tilde U = 4\pi\tilde\sigma\delta(\zeta)
1132: \end{equation}
1133: with the boundary condition
1134: \begin{equation}\label{N2}
1135: \left.\tilde U_{,\tilde\zeta}\right|_{\tilde\zeta=0+}=2\pi\tilde\sigma,
1136: \end{equation}
1137: where $\tilde\sigma=\tilde\sigma(\tilde\rho)$ is the surface mass
1138: density of the final disk. With $\dd M=2\pi\sigma\dd\rho$ and $\dd
1139: J=\Omega\rho^2\dd M$, Eq.~\eqref{24} leads to the additional boundary
1140: conditions
1141: \begin{equation}\label{N3}
1142: \tilde\sigma(\tilde\rho)=
1143: 2\sigma(\rho)\frac{\rho}{\tilde\rho}\frac{\dd\rho}{\dd\tilde\rho},\quad
1144: \tilde\Omega(\tilde\rho)\tilde\rho^2=\Omega_0\rho^2.
1145: \end{equation}
1146: The initial surface mass density of the Maclaurin disk is
1147: \begin{equation}\label{N4}
1148: \sigma(\rho)=\frac{3M}{2\pi\rho_0^2}\sqrt{1-\frac{\rho^2}{\rho_0^2}}
1149: \end{equation}
1150: and the initial constant angular velocity $\Omega_0$ is related to the
1151: initial mass by
1152: \begin{equation}\label{N5}
1153: \Omega_0^2=\frac{3\pi M}{4\rho_0^3}.
1154: \end{equation}
1155: Using these relations, together with the Euler equation
1156: \begin{equation}
1157: \left.\tilde U_{,\tilde\rho}\right|_{\tilde\zeta=0}=\tilde\Omega^2(\tilde\rho)
1158: \tilde\rho,
1159: \end{equation}
1160: we find that a (rigidly rotating) Maclaurin disk with the parameters
1161: \begin{equation}\label{N6}
1162: \tilde\Omega=4\Omega_0,\quad \tilde\rho_0=\frac{1}{2}\rho_0
1163: \end{equation}
1164: indeed solves the boundary value problem \eqref{N1}-\eqref{N3}.
1165:
1166:
1167:
1168: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1169: \section{Discussion}
1170:
1171: In this paper we have performed the analysis
1172: of collision processes in the spirit of equilibrium thermodynamics.
1173: Avoiding the solution
1174: of the
1175: full dynamical problem, we compared initial and final
1176: equilibrium configurations to obtain a \lq\lq rough\rq\rq\
1177: picture of these processes. In this way we were able to calculate the
1178: energy loss by the emission of gravitational waves and to find conditions
1179: (\lq\lq parameter relations\rq\rq) for the formation of final stars and disks.
1180:
1181: The application of this method to
1182: collisions of perfect fluid stars and collisions of
1183: rigidly rotating disks of dust leads to restrictions of the
1184: initial parameters.
1185: It turned out that the formation of final stars/disks from stars/disks
1186: is only possible
1187: for a subset of the parameter space of the
1188: initial objects. Otherwise, the collision of spheres and disks
1189: would lead to other final
1190: states, e.g. to black holes.
1191:
1192: Our main result is the numerical solution of the Einstein equations for the
1193: differentially rotating (DR) disk formed
1194: by the collision of two identical rigidly rotating (RR)
1195: disks with parallel angular
1196: momenta. We calculated the characteristic quantities of the final
1197: DR disk, as for
1198: example the rotation curve $\tilde\Omega(\tilde\rho)$ as it depends on
1199: the centrifugal
1200: parameter $\mu$ of the initial RR disks. It turned out, that the angular
1201: velocity $\tilde\Omega$
1202: is almost constant (as shown in Sec.~\ref{Newtonian}, it is
1203: \emph{strictly} constant in the Newtonian limit). Therefore, the simplified
1204: model of the formation of an RR disk from the
1205: collision of two RR
1206: disks as presented in
1207: %\cite{Hennig}
1208: Hennig \& Neugebauer (2006),
1209: which has to allow frictional processes to reach constant angular velocity,
1210: turns out to be a good approximation to our present purely gravitational
1211: (frictionless) model (B).
1212:
1213: \begin{table}[t]\centering
1214: \begin{tabular}{lr}
1215: \hline
1216: {\it colliding objects} & $\eta_\textrm{max}$\\
1217: \hline\hline
1218: Schwarzschild BHs & 29.3\%\\
1219: RR disks & 23.8\%\\
1220: Schwarzschild stars \quad & 19.7\%\\
1221: Neutron stars & 2.3\%\\
1222: \hline
1223: \end{tabular}
1224: \caption{Upper limits for the efficiency $\eta$ of different collision
1225: processes including Hawking's and Ellis' limit for the collision of two
1226: spherically symmetric black holes \cite{Hawking}.
1227: According to the last plot of Fig.~\ref{fig_DR}
1228: [$\eta^{\textrm{F}}(1.954\dots)=0$] the two efficiencies
1229: $\eta^{\textrm{RR}}$ and $\eta^{\textrm{DR}}$ coincide with a maximum
1230: value
1231: $\eta^{\textrm{RR}}_{\textrm{max}}=\eta^{\textrm{DR}}_{\textrm{max}}\approx
1232: 23.8\%$.
1233: }
1234: \label{table}
1235: \end{table}
1236:
1237: For each of the studied collision scenarios, we calculated an upper limit
1238: for the energy of the emitted gravitational waves. A summary of the maximal
1239: efficiencies is given in table~\ref{table}. The value
1240: $\eta_{\textrm{max}}\approx 2.3\%$ for the collision of Neutron stars is
1241: relatively small compared to the other examples. The reason is the
1242: restricted equation of state (completely degenerate ideal Fermi gas)
1243: that does not allow for strongly relativistic stars.
1244:
1245:
1246:
1247:
1248: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1249: \begin{acknowledgments}
1250: We would like to thank David Petroff for many valuable discussions.
1251: This work was supported by the Deutsche Forschungsgemeinschaft (DFG)
1252: through the SFB/TR7 \lq\lq Gravitationswellenastronomie\rq\rq.
1253: \end{acknowledgments}
1254:
1255:
1256: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1257: \appendix
1258:
1259: \section{Potentials of the rigidly rotating disk of dust}\label{AppendixA}
1260:
1261: For the numerical calculation of the DR disk that is formed by the
1262: collision of two RR disks we need some formulae for
1263: quantities of the RR disk of dust.
1264:
1265: The coefficient $V_0$ in the four-velocity \eqref{22a} as a
1266: function of the parameter $\mu$ can be calculated from a
1267: very rapidly converging
1268: series, cf. \cite{Kleinwaechter},
1269: \begin{eqnarray}\label{A1}
1270: \coth\frac{V_0}{2} & = & -\frac{4}{\mu}+
1271: 0.0294938052100425142\mu+5.4681333461446\cdot10^{-6}\mu^3\nonumber\\
1272: && -1.07467432587\cdot10^{-9}\mu^5+2.1127368\cdot10^{-13}\mu^7\\
1273: && -4.154\cdot10^{-17}\mu^9+\mathcal O(\mu^{11}).\nonumber
1274: \end{eqnarray}
1275: The disk values ($\zeta=0$, $\rho\le\rho_0$)
1276: of the metric functions $U$ and $a$ and the mass density $\sigma$ are given
1277: by the equations
1278: \begin{equation}\label{A2}
1279: \ee^{2U}=\ee^{2V_0(\hat\mu)}-\frac{\mu\rho^2}{2\rho_0^2},
1280: \end{equation}
1281: \begin{equation}\label{A3}
1282: (1+\Omega_0 a)\ee^{2U}=\ee^{V_0(\mu)}\ee^{V_0(\hat\mu)},
1283: \end{equation}
1284: \begin{equation}\label{A4}
1285: \sigma=-\frac{\Omega_0}{2\pi\ee^{V_0(\mu)}}
1286: \frac{b_0'(\hat\mu)}
1287: {\ee^{V_0\hat\mu}},
1288: \end{equation}
1289: with
1290: \begin{equation}\label{A5}
1291: \Omega_0=\sqrt{\frac{\mu}{2}}\frac{\ee^{V_0}}{\rho_0},\quad
1292: b_0=-\sqrt{1-\ee^{4V_0}-4\Omega_0^2\rho_0^2},
1293: \end{equation}
1294: cf. \cite{Neugebauer2}.
1295: The notation $V_0(\hat\mu)$, $b'_0(\hat\mu)$ indicates that the argument
1296: $\mu$ in the parameter functions $V_0(\mu)$ and $b'_0(\mu)$ has to be
1297: replaced by $\hat\mu=(1-\rho^2/\rho_0^2)\mu$. $b'_0(\hat\mu)$ means
1298: $\dd b_0(\hat\mu)/\dd\hat\mu$.
1299:
1300:
1301:
1302:
1303:
1304:
1305: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1306:
1307:
1308:
1309:
1310: \begin{thebibliography}{}
1311:
1312: \bibitem [Anninos et al. 1993]
1313: {Anninos1} Anninos, P., Hobill, D., Seidel, E., Smarr, L., \&
1314: Suen, W.-M. 1993, Phys. Rev. Lett., 71, 2851
1315:
1316: \bibitem[Anninos et al. 1995]
1317: {Anninos2} Anninos, P., Hobill, D., Seidel, E., Smarr, L., \&
1318: Suen, W.-M. 1995, Phys. Rev. Lett., 52, 2044
1319:
1320: \bibitem[Anninos et al. 1998]
1321: {Anninos3} Anninos, P., \& Brandt, S. 1998, Phys. Rev. Lett., 81, 508
1322:
1323: \bibitem[Ansorg \& Meinel 2000]
1324: {Ansorg1} Ansorg, M., \& Meinel, R. 2000, Gen. Rel. Grav., 32, 1365
1325:
1326: \bibitem[Ansorg 2001]
1327: {Ansorg2} Ansorg, M. 2001, Gen. Rel. Grav., 33, 309
1328:
1329: \bibitem[Ansorg \& Petroff 2005]
1330: {Rings} Ansorg, M., \& Petroff, D. 2005, Phys. Rev. D, 72, 024019
1331:
1332: \bibitem[Bardeen \& Wagoner 1969]
1333: {Bardeen1} Bardeen, J. M., \& Wagoner, R. V. 1969, ApJ, 158, L65
1334:
1335: \bibitem[Bardeen \& Wagoner 1971]
1336: {Bardeen2} Bardeen, J. M., \& Wagoner, R. V. 1971, ApJ, 167, 359
1337:
1338: \bibitem[Bardeen 1973]
1339: {BlackHoles} Bardeen, J. M. 1973, in {\it Black Holes, Les astres
1340: occlus}, ed. DeWitt, C. \& DeWitt, B., (Gordon and Breach Science
1341: Publishers, New York)
1342:
1343: \bibitem[Beig \& Simon 1992]
1344: {Beig} Beig, R., \& Simon, W. 1992, Commun. Math. Phys., 144, 373
1345:
1346: \bibitem[Eppley 1975]
1347: {Eppley} Eppley, K. 1975, Ph.D. thesis, Princeton University
1348:
1349: \bibitem[Fiske et al. 2005]
1350: {Fiske} Fiske, D. R., Baker, J. G., van Meter, R. R., Choi, \& D.-I.,
1351: Centrella, J. M. 2005, Phys. Rev. D, 71, 104036
1352:
1353: \bibitem[Hawking \& Ellis 1973]
1354: {Hawking} Hawking, S. W., \& G. F. R. Ellis, G. F. R. 1973,
1355: The large scale
1356: structure of space-time, Cambridge Monographs on Mathematical
1357: Physics, (Cambridge University Press)
1358:
1359: \bibitem[Hennig \& Neugebauer 2006]
1360: {Hennig} Hennig, J., \& Neugebauer, G. 2006, Phys. Rev. D, 74, 064025
1361:
1362: \bibitem[Kleinw\"achter 1995]
1363: {Kleinwaechter} Kleinw\"achter, A. 1995, Ph.D. thesis,
1364: Friedrich-Schiller-Universit\"at Jena
1365:
1366: \bibitem[Lindblom \& Masood-ul-Alam 1994]
1367: {Lindblom} Lindblom, L., \& Masood-ul-Alam, A. K. M. 1994,
1368: Commun. Math. Phys., 162, 123
1369:
1370: \bibitem[L\"offler et al. 2006]
1371: {Loeffler} L\"offler, F., Rezzolla, L., \& Ansorg, M. 2006,
1372: Phys. Rev. D, 74, 104018
1373:
1374: \bibitem[Masood-ul-Alam 2007]{Masood}
1375: Masood-ul-Alam, A. K. M. 2007, Gen. Rel. Grav., 39, 55
1376:
1377: \bibitem[Misner et al. 2002]
1378: {MTW} Misner, C. W., Thorne, K. S., \& Wheeler, J. A 2002, Gravitation,
1379: (W. H. Freeman and Company, New York)
1380:
1381: \bibitem[Neugebauer \& Meinel 1993]
1382: {Neugebauer1} Neugebauer, G., \& Meinel, R. 1993, ApJ., 414, L97
1383:
1384: \bibitem[Neugebauer \& Meinel 1994]
1385: {Neugebauer2} Neugebauer, G., \& Meinel, R. 1994, Phys. Rev. Lett.,
1386: 73, 2166
1387:
1388: \bibitem[Neugebauer \& Meinel 1995]
1389: {Neugebauer3} Neugebauer, G., \& Meinel, R. 1995, Phys. Rev. Lett.,
1390: 75, 3046
1391:
1392: \bibitem[Neugebauer et al. 1996]
1393: {Neugebauer4} Neugebauer, G., Kleinw\"achter, A., \& Meinel, R. 1996,
1394: Helv. Phys. Acta, 69, 472
1395:
1396: \bibitem[Shapiro \& Teukolsky 1983]
1397: {Shapiro} Shapiro, S. L., \& Teukolsky, S. A., Black Holes,
1398: White Dwarfs, and Neutron Stars --- The Physics of Compact Objects,
1399: (John Wiley \& Sons, New York)
1400:
1401: \bibitem[Smarr et al. 1976]
1402: {Smarr} Smarr, L., {\v C}ade{\v z}, A., DeWitt, B., \& Eppley, K. 1976,
1403: Phys. Rev. D, 14, 2443
1404:
1405: \bibitem[Sperhake et al. 2005]
1406: {Sperhake1} Sperhake, U., Kelly, B. , Laguna, P., Smith, K. L., \&
1407: Schnetter, E. 2005, Phys. Rev. D, 71, 124042
1408:
1409: \bibitem[Sperhake 2006]
1410: {Sperhake2} Sperhake, U. 2006, gr-qc/0606079
1411:
1412: \bibitem[Zlochower et al. 2005]
1413: {Zlochower} Zlochower, Y., Baker, J. G., Campanelli, M.,
1414: Lousto, C. O. 2005, Phys. Rev. D, 72, 024021
1415:
1416:
1417: \end{thebibliography}
1418:
1419: \end{document}
1420:
1421: