1: % Template article for preprint document class `elsart'
2: % SP 2001/01/05
3: % Modified CG (ESME) for Model 4, single column, 2 titles, abstract/rÈsumÈ,
4: % and 2 sets of keywords - 21.05.03 - file called phys-english.tex
5: % English Version for Physique (CRAS series 4, now COMREN) - a Note in English
6: % Revamped, CG, 18.08.04, adding header, dates,
7: % and name of presenter
8:
9: \documentclass{elsart4-1}
10:
11: % Use the option doublespacing or reviewcopy to obtain double line spacing
12: % \documentclass[doublespacing]{elsart}
13:
14: % if you use PostScript figures in your article
15: % use the graphics package for simple commands
16: % \usepackage{graphics}
17: % or use the graphicx package for more complicated commands
18: % \usepackage{graphicx}
19: % or use the epsfig package if you prefer to use the old commands
20: % \usepackage{epsfig}
21:
22: \bibliographystyle{unsrt}
23:
24: %\allowdisplaybreaks
25:
26: % The amssymb package provides various useful mathematical symbols
27: \usepackage{amssymb}
28: \usepackage{amsfonts}
29: \usepackage{amsmath}
30:
31: \usepackage[english,francais]{babel}
32:
33: %\usepackage{showkeys}
34:
35: %ENVIRONMENTS THEOREMS...
36: % These are predefined, and correspond to the
37: % numbering system used in the journal!
38: %English
39: \newtheorem{theorem}{Theorem}[section]
40: \newtheorem{lemma}[theorem]{Lemma}
41: \newtheorem{e-proposition}[theorem]{Proposition}
42: \newtheorem{corollary}[theorem]{Corollary}
43: \newtheorem{e-definition}[theorem]{Definition\rm}
44: \newtheorem{remark}{\it Remark\/}
45: \newtheorem{example}{\it Example\/}
46: %French
47: \newtheorem{theoreme}{Th\'eor\`eme}[section]
48: \newtheorem{lemme}[theoreme]{Lemme}
49: \newtheorem{proposition}[theoreme]{Proposition}
50: \newtheorem{corollaire}[theoreme]{Corollaire}
51: \newtheorem{definition}[theoreme]{D\'efinition\rm}
52: \newtheorem{remarque}{\it Remarque}
53: \newtheorem{exemple}{\it Exemple\/}
54: \renewcommand{\theequation}{\arabic{equation}}
55: \setcounter{equation}{0}
56:
57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
58: %% GUILLEMETS (FRENCH QUOTES) %%
59: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
60: \def\og{\leavevmode\raise.3ex\hbox{$\scriptscriptstyle\langle\!\langle$~}}
61: \def\fg{\leavevmode\raise.3ex\hbox{~$\!\scriptscriptstyle\,\rangle\!\rangle$}}
62:
63: \begin{document}
64: % Select a primary header Physics or Astrophysics
65: % You can place after the header (classification), if you know it.
66:
67: \centerline{Physics or Astrophysics/Header}
68: \begin{frontmatter}
69:
70: % Title, authors and addresses
71:
72: % use the thanksref command within \title, \author or \address for footnotes;
73: % use the ead command for the email address,
74: % and the form \ead[url] for the home page:
75: % \title{Title\thanksref{label1}}
76: % \thanks[label1]{}
77: % \author{Name\thanksref{label2}}
78: % \ead{email address}
79: % \ead[url]{home page}
80: % \thanks[label2]{}
81: % \address{Address\thanksref{label3}}
82: % \thanks[label3]{}
83: \selectlanguage{english} \title{General Relativistic Dynamics of Compact
84: Binary Systems}
85:
86: % use optional labels to link authors explicitly to addresses:
87: % \author[label1,label2]{}
88: % \address[label1]{}
89: % \address[label2]{}
90: % If all authors are at the same address, the [label1] can be suppressed
91:
92: \selectlanguage{english}
93: \author[authorlabel1]{Luc Blanchet},
94: \ead{blanchet@iap.fr}
95:
96: \address[authorlabel1]{${\mathcal{G}}{\mathbb{R}}
97: \varepsilon{\mathbb{C}}{\mathcal{O}}$, Institut d'Astrophysique de
98: Paris, \\ UMR 7095 CNRS, Universit\'e Pierre \& Marie Curie, \\ 98 bis
99: boulevard Arago, 75014 Paris, France}
100:
101: % If your know the dates of reception, and acceptation you can put them now;
102: % idem the name of the person presenting your article
103:
104: \medskip
105: \begin{center}
106: %{\small Received *****; accepted after revision +++++}
107: \end{center}
108:
109: \begin{abstract}
110: The equations of motion of compact binary systems have been derived in
111: the post-Newtonian (PN) approximation of general relativity. The current
112: level of accuracy is 3.5PN order. The conservative part of the equations
113: of motion (neglecting the radiation reaction damping terms) is deducible
114: from a generalized Lagrangian in harmonic coordinates, or equivalently
115: from an ordinary Hamiltonian in ADM coordinates. As an application we
116: investigate the problem of the dynamical stability of circular binary
117: orbits against gravitational perturbations up to the 3PN order. We find
118: that there is no innermost stable circular orbit or ISCO at the 3PN
119: order for equal masses.
120:
121: %{\it To cite this
122: %article: A. Name1, A. Name2, C. R. Physique 6 (2005).}
123:
124: \vskip 0.5\baselineskip
125:
126: \selectlanguage{francais}
127: \noindent{\bf R\'esum\'e}
128: \vskip 0.5\baselineskip
129: \noindent
130: Les \'equations du mouvement d'un syst\`eme binaire d'objets compacts
131: ont \'et\'ees calcul\'ees dans l'approximation post-newtonienne (PN) de
132: la relativit\'e g\'en\'erale. Le niveau d'approximation atteint l'ordre
133: 3.5PN. La partie conservative des \'equations du mouvement (obtenue en
134: n\'egligeant les termes de freinage de rayonnement) se d\'eduit d'un
135: lagrangien g\'en\'eralis\'e en coordonn\'ees harmoniques, et, de
136: fa\c{c}on \'equivalente, d'un hamiltonien ordinaire en coordonn\'ees
137: ADM. Comme application nous \'etudions le probl\`eme de la stabilit\'e
138: dynamique, vis-\`a-vis des perturbations gravitationnelles, des orbites
139: binaires circulaires \`a l'ordre 3PN. Nous trouvons qu'il n'y a pas de
140: derni\`ere orbite stable circulaire ou ISCO \`a l'ordre 3PN pour des
141: masses \'egales.
142:
143: %{\it Pour citer cet article~: A. Name1, A. Name2, C. R.
144: %Physique 6 (2005).}
145:
146: %Now keywords/mots-cl\`Es
147: \keyword{Post-Newtonian theory; equations of motion; compact binary
148: systems} \vskip 0.5\baselineskip
149: \noindent{\small{\it Mots-cl\'es~:} Th\'eorie post-newtonienne;
150: \'equations du mouvement; syst\`emes binaires compacts}}
151: \end{abstract}
152: \end{frontmatter}
153:
154: % now the Version franÁaise abr\`Eg\`Ee, if it exists
155: %\selectlanguage{francais}
156: %\section*{Version fran\c{c}aise abr\'eg\'ee}
157: % Text of your Version franÁaise abrègèe here
158:
159: \selectlanguage{english}
160: % main text
161:
162: \section{Introduction}\label{secI}
163:
164: The problem of the dynamics of two compact bodies is part of a program
165: aimed at unravelling the information contained in the
166: gravitational-wave signals emitted by inspiralling and coalescing
167: compact binaries --- systems of neutron stars or black holes driven
168: into coalescence by emission of gravitational radiation. The treatment
169: of the problem is Post-Newtonian, \textit{i.e.} based on formal
170: expansions, when the speed of light $c$ tends to infinity, of general
171: relativity theory around Newton's theory. The early, classic works of
172: Lorentz \& Droste~\cite{LD17}, Eddington \& Clark~\cite{EC38},
173: Einstein, Infeld \& Hoffmann~\cite{EIH}, Fock~\cite{Fock},
174: Papapetrou~\cite{Papa51} and others led to a good understanding of the
175: equations of motion of $N$ bodies at the first post-Newtonian
176: approximation (1PN, corresponding to order~$\sim 1/c^2$). In the
177: 1970's, a series of works~\cite{OO73,OO74a,OO74b} led to a nearly
178: complete control of the problem of motion at the second post-Newtonian
179: approximation (2PN~$\sim 1/c^4$). Then, starting in the early 80's,
180: motivated by the observation of secular orbital effects in the
181: Hulse-Taylor binary pulsar PSR1913+16~\cite{TFMc79,T93}, several
182: groups solved the two-body problem at the 2.5PN level (while
183: completing on the way the derivation of the 2PN
184: dynamics)~\cite{BeDD81,DD81a,DD81b,D83houches,DS85,GKop86,BFP98,IFA01}. The
185: 2.5PN term constitutes the first contribution of gravitational
186: \textit{radiation reaction} in the equations of motion (such term is
187: the analogue of the Abraham--Lorentz reaction force in
188: electromagnetism), and is directly responsible for the decrease of the
189: binary pulsar orbital period by emission of gravitational radiation.
190:
191: In the late 90's, motivated by the aim of deriving high-accuracy
192: templates for the data analysis of the international network of
193: interferometric gravitational-wave detectors LIGO/VIRGO, two groups
194: embarked on the derivation of the equations of motion at the third
195: post-Newtonian (3PN) level. One group used the Arnowitt-Deser-Misner
196: (ADM) Hamiltonian formalism of general
197: relativity~\cite{JaraS98,JaraS99,DJSpoinc,DJSequiv} and worked in a
198: corresponding ADM-type coordinate system. Another group used a direct
199: post-Newtonian iteration of the equations of motion in harmonic
200: coordinates~\cite{BF00,BFeom,BFreg,BFregM,ABF01,BI03CM}. The end
201: results of these two approaches have been proved to be physically
202: equivalent~\cite{DJSequiv,ABF01}. However, both approaches, even after
203: exploiting all symmetries and pushing their
204: Hadamard-regularization-based methods to the maximum, left
205: undetermined one dimensionless parameter at the 3PN order. The
206: appearance of this unknown parameter was related with the choice of
207: the regularization method used to cure the self-field divergencies of
208: point particles. Both lines of works regularized the self-field
209: divergencies by some version of the Hadamard regularization
210: method. Finally, the completion of the equations of motion at the 3PN
211: order was made possible thanks to the powerful \textit{dimensional}
212: self-field regularization, which could fix up uniquely the value of
213: the ambiguity parameter~\cite{DJSdim,BDE04}. This result is also in
214: complete agreement with the recent finding of~\cite{itoh1,itoh2}, who
215: derived the 3PN equations of motion in harmonic gauge using a
216: ``surface-integral'' approach without use of self-field
217: regularization. Finally, the 3.5PN terms, which constitute a 1PN
218: relative modification of the radiation reaction force (and are
219: relatively easier to derive), have been added in
220: Refs.~\cite{IW93,IW95,JaraS97,PW02,KFS03,NB05}.
221:
222: In Section~\ref{secII} of the present paper we give the final result for
223: the equations of motion of compact binaries at the 3.5PN order. The
224: equations are presented in ready to use quasi-Newtonian form, in the
225: reference frame associated with the center of mass position. In
226: Section~\ref{secIII} we discuss the Lagrangian and Hamiltonian
227: formulations of the conservative part of the equations, obtained by
228: neglecting the radiation reaction terms occuring at the 2.5PN and 3.5PN
229: orders. Finally, in Section~\ref{secIV}, we investigate, following
230: Ref.~\cite{BI03CM}, the question of the stability of circular orbits,
231: against linear gravitational perturbations, up to the 3PN order.
232:
233: \section{Equations of motion in the center-of-mass frame}\label{secII}
234: %
235: In the present paper we employ the so-called harmonic-coordinates
236: approach~\cite{BF00,BFeom,ABF01,BI03CM} which derived the 3PN binary's
237: equations of motion in harmonic coordinates. From these equations,
238: obtained at first in a general frame, we translate the origin of
239: coordinates to the binary's center-of-mass by imposing that the
240: binary's center-of-mass vector position is ${\rm G}_i = 0$. The
241: center-of-mass vector is nothing but the conserved integral of the
242: motion that is associated, {\it via} the Noether theorem, with the
243: \textit{boost symmetry} of the Lagrangian from which the 3PN equations
244: of motion are derived~\cite{ABF01,BI03CM}. The condition ${\rm G}_i =
245: 0$ results in the 3PN-accurate relationship between the individual
246: positions of the particles $y_1^i$ and $y_2^i$ in the center-of-mass
247: frame, and the relative position $x^i= y_{1}^i-y_{2}^i$ and velocity
248: $v^i= v_{1}^i-v_{2}^i = d x^i/dt$. The center-of-mass equations of
249: motion are then obtained by replacing the individual positions and
250: velocities by their center-of-mass expressions, given in terms of
251: $x^i$ and velocity $v^i$, applying as usual the order-reduction of all
252: accelerations where necessary. Order reduction means that any
253: acceleration (or derivative of acceleration) in a sub-dominant
254: post-Newtonian term is to be replaced by its explicit expression given
255: as a function of the positions and velocities as deduced from the
256: lower-order equations of motion themselves.
257:
258: We shall denote the orbital separation by $r=\vert{\bf x}\vert$, and
259: pose ${\bf n}={\bf x}/r$ and $\dot{r}={\bf n}\cdot{\bf v}$. The mass
260: parameters are the total mass $m=m_1+m_2$, the mass difference $\delta
261: m=m_1-m_2$, the reduced mass $\mu=m_1m_2/m$, and the very useful
262: symmetric mass ratio $\nu=\mu/m$, which is such that $0<\nu\leq 1/4$,
263: with $\nu=1/4$ in the case of equal masses, and $\nu\to 0$ in the
264: ``test-mass'' limit for one of the bodies. We write the relative
265: acceleration in the center-of-mass frame in the form (we pose $G=1$)
266: %
267: \begin{equation}\label{eom}
268: \frac{d v^i}{dt}=-\frac{m}{r^2}\Big[(1+\mathcal{A})\,n^i +
269: \mathcal{B}\,v^i \Big]+ \mathcal{O}\left( \frac{1}{c^8} \right)\;,
270: \end{equation}
271: %
272: where the first term represents the famous Newtonian approximation, and
273: where the post-Newtonian remainder term $\mathcal{O}( c^{-8})$ indicates
274: the level of accuracy of the expression which is here 3.5PN order. We
275: find~\cite{BI03CM,NB05} that the coefficients $\mathcal{A}$ and
276: $\mathcal{B}$ are
277: %
278: \allowdisplaybreaks{\begin{eqnarray}
279: \mathcal{A}&=& \frac{1}{c^2}\left\{-\frac{3\,\dot{r}^2\,\nu}{2} + v^2 +
280: 3\,\nu\,v^2-\frac{m}{r}\left(4 +2\,\nu \right)\right\}\nonumber\\ &+&
281: \frac{1}{c^4}\left\{\frac{15\,\dot{r}^4\,\nu}{8} -
282: \frac{45\,\dot{r}^4\,\nu^2}{8} - \frac{9\,\dot{r}^2\,\nu\,v^2}{2} +
283: 6\,\dot{r}^2\,\nu^2\,v^2 + 3\,\nu\,v^4 - 4\,\nu^2\,v^4\right.
284: \nonumber\\ &&\qquad + \left.\frac{m}{r}\left( -2\,\dot{r}^2 -
285: 25\,\dot{r}^2\,\nu - 2\,\dot{r}^2\,\nu^2 - \frac{13\,\nu\,v^2}{2} +
286: 2\,\nu^2\,v^2 \right)\right. \nonumber\\ &&\qquad +\left.\frac{m^2}{r^2}\,\left( 9 + \frac{87\,\nu}{4}
287: \right)\right\}\nonumber\\&+&\frac{1}{c^5}\left\{-
288: \frac{24\,\dot{r}\,\nu\,v^2}{5}\frac{m}{r}-\frac{136\,\dot{r}\,\nu}{15} \frac{m^2}{r^2}\right\}\nonumber\\ &+&
289: \frac{1}{c^6}\left\{-\frac{35\,\dot{r}^6\,\nu}{16} +
290: \frac{175\,\dot{r}^6\,\nu^2}{16} -
291: \frac{175\,\dot{r}^6\,\nu^3}{16}+\frac{15\,\dot{r}^4\,\nu\,v^2}{2}
292: \right.\nonumber\\&&\qquad - \left. \frac{135\,\dot{r}^4\,\nu^2\,v^2}{4}
293: + \frac{255\,\dot{r}^4\,\nu^3\,v^2}{8} -
294: \frac{15\,\dot{r}^2\,\nu\,v^4}{2} + \frac{237\,\dot{r}^2\,\nu^2\,v^4}{8}
295: \right.\nonumber\\ &&\qquad -\left. \frac{45\,\dot{r}^2\,\nu^3\,v^4}{2}
296: + \frac{11\,\nu\,v^6}{4} - \frac{49\,\nu^2\,v^6}{4} + 13\,\nu^3\,v^6
297: \right.\nonumber\\ &&\qquad + \left.\frac{m}{r}\left(
298: 79\,\dot{r}^4\,\nu - \frac{69\,\dot{r}^4\,\nu^2}{2} -
299: 30\,\dot{r}^4\,\nu^3 - 121\,\dot{r}^2\,\nu\,v^2 +
300: 16\,\dot{r}^2\,\nu^2\,v^2 \right.\right.\nonumber\\&&\qquad\qquad\quad~
301: +\left.\left. 20\,\dot{r}^2\,\nu^3\,v^2+\frac{75\,\nu\,v^4}{4} +
302: 8\,\nu^2\,v^4 - 10\,\nu^3\,v^4 \right)\right.\nonumber\\ &&\qquad +
303: \left. \frac{m^2}{r^2}\,\left( \dot{r}^2 +
304: \frac{32573\,\dot{r}^2\,\nu}{168} + \frac{11\,\dot{r}^2\,\nu^2}{8} -
305: 7\,\dot{r}^2\,\nu^3 + \frac{615\,\dot{r}^2\,\nu\,\pi^2}{64} -
306: \frac{26987\,\nu\,v^2}{840}
307: \right.\right.\nonumber\\&&\qquad\qquad\quad~ +\left.\left. \nu^3\,v^2 -
308: \frac{123\,\nu\,\pi^2\,v^2}{64} - 110\,\dot{r}^2\,\nu\,\ln
309: \Big(\frac{r}{r'_0}\Big) + 22\,\nu\,v^2\,\ln \Big(\frac{r}{r'_0}\Big)
310: \right)\right.\nonumber\\&&\qquad +\left.\frac{m^3}{r^3}\left( -16 -
311: \frac{437\,\nu}{4} - \frac{71\,\nu^2}{2} + \frac{41\,\nu\,{\pi }^2}{16}
312: \right)\right\}\nonumber\\ &+& \frac{1}{c^7}\left\{ \frac{m}{r} \,
313: \left( \frac{366}{35}\,\nu\,v^4 + 12 \nu^2\,v^4 - 114\,v^2\,\nu
314: \dot{r}^2 - 12 \nu^2\,v^2 \dot{r}^2 + 112 \nu\,\dot{r}^4\right) \right.
315: \nonumber\\ & & \qquad\left. + \frac{m^2}{r^2} \, \left(
316: \frac{692}{35}\,\nu\,v^2 - \frac{724}{15}\,v^2\, \nu^2 +
317: \frac{294}{5}\,\nu\,\dot{r}^2 + \frac{376}{5} \nu^2\,\dot{r}^2
318: \right)\right. \nonumber\\ & & \qquad\left. + \frac{m^3}{r^3} \,
319: \left( \frac{3956}{35}\nu + \frac{184}{5} \nu^2 \right)
320: \right\},\nonumber\\ \mathcal{B}&=&\frac{1}{c^2}\Big\{-4\,\dot{r} +
321: 2\,\dot{r}\,\nu\Big\}
322: \nonumber\\&+&\frac{1}{c^4}\left\{\frac{9\,\dot{r}^3\,\nu}{2} +
323: 3\,\dot{r}^3\,\nu^2 -\frac{15\,\dot{r}\,\nu\,v^2}{2} -
324: 2\,\dot{r}\,\nu^2\,v^2\right.\nonumber\\ &&\qquad + \left.\frac{m}{r}\left( 2\,\dot{r} + \frac{41\,\dot{r}\,\nu}{2} + 4\,\dot{r}\,\nu^2
325: \right)\right\}\nonumber\\&+&\frac{1}{c^5}\left\{
326: \frac{8\,\nu\,v^2}{5}\frac{m}{r}+\frac{24\,\nu}{5} \frac{m^2}{r^2}\right\}\nonumber\\ &+&
327: \frac{1}{c^6}\left\{-\frac{45\,\dot{r}^5\,\nu}{8} + 15\,\dot{r}^5\,\nu^2
328: + \frac{15\,\dot{r}^5\,\nu^3}{4} + 12\,\dot{r}^3\,\nu\,v^2
329: \right.\nonumber\\&&\qquad -\left. \frac{111\,\dot{r}^3\,\nu^2\,v^2}{4}
330: -12\,\dot{r}^3\,\nu^3\,v^2 -\frac{65\,\dot{r}\,\nu\,v^4}{8} +
331: 19\,\dot{r}\,\nu^2\,v^4 + 6\,\dot{r}\,\nu^3\,v^4
332: \right.\nonumber\\&&\qquad\left. + \frac{m}{r}\left(
333: \frac{329\,\dot{r}^3\,\nu}{6} + \frac{59\,\dot{r}^3\,\nu^2}{2} +
334: 18\,\dot{r}^3\,\nu^3 - 15\,\dot{r}\,\nu\,v^2 - 27\,\dot{r}\,\nu^2\,v^2 -
335: 10\,\dot{r}\,\nu^3\,v^2 \right) \right.\nonumber\\&&\qquad
336: +\left.\frac{m^2}{r^2}\,\left( -4\,\dot{r} -
337: \frac{18169\,\dot{r}\,\nu}{840} + 25\,\dot{r}\,\nu^2 + 8\,\dot{r}\,\nu^3
338: - \frac{123\,\dot{r}\,\nu\,\pi^2}{32} + 44\,\dot{r}\,\nu\,\ln
339: \Big(\frac{r}{r'_0}\Big) \right)\right\}\nonumber\\ &+&
340: \frac{1}{c^7}\left\{ \frac{m}{r} \, \left( - \frac{626}{35}\,\nu\,v^4
341: - \frac{12}{5} \nu^2\,v^4 + \frac{678}{5}\,\nu\,v^2 \dot{r}^2 +
342: \frac{12}{5} \nu^2\,v^2 \dot{r}^2 - 120 \nu\,\dot{r}^4\right)
343: \right.\nonumber\\ & & \qquad\left. + \frac{m^2}{r^2} \, \left(
344: \frac{164}{21}\,\nu\,v^2 + \frac{148}{5} \nu^2\,v^2 -
345: \frac{82}{3}\,\nu\,\dot{r}^2 - \frac{848}{15} \nu^2\,\dot{r}^2 \right)
346: \right.\nonumber\\ & & \qquad\left. + \frac{m^3}{r^3} \, \left( -
347: \frac{1060}{21}\nu - \frac{104}{5}\nu^2 \right)\right\}.\label{ABcoeff}
348: \end{eqnarray}}\noindent
349: %
350: The 3.5PN equations of motion play a crucial role when deriving the
351: high-accuracy templates which will be used for analysing (hopefully in a
352: near future) the gravitational wave signals from compact binary inspiral
353: in the data analysis of the LIGO and VIRGO detectors.
354:
355: At the 3PN order we find some logarithmic terms, depending on some
356: arbitrary constant $r'_0$. The presence of these logarithms reflects in
357: fact the use of a specific harmonic coordinate system. It is indeed
358: known that the logarithms at the 3PN order in Eqs.~(\ref{ABcoeff}),
359: together with the constant $r'_0$ therein, can be removed by applying a
360: gauge transformation. This shows that there is no physics associated
361: with them, and that these logarithms and the constant $r'_0$ will never
362: appear in any physical result derived from these equations, because the
363: physical results must be gauge invariant ($r'_0$ is sometimes referred
364: to as a ``gauge constant''). The gauge transformation at 3PN order whose
365: effect is to remove the logarithms is given in~\cite{BFeom}. Notice that
366: after applying this gauge transformation we are still within the class
367: of harmonic coordinates. The resulting modification of the equations of
368: motion affects only the coefficients of the 3PN order in
369: Eqs.~(\ref{ABcoeff}); let us denote them by $\mathcal{A}_{\rm 3PN}$ and
370: $\mathcal{B}_{\rm 3PN}$. The new values of these coefficients, say
371: $\mathcal{A}'_{\rm 3PN}$ and $\mathcal{B}'_{\rm 3PN}$, obtained after
372: removal of the logarithms by the latter harmonic gauge transformation,
373: are then~\cite{MW03}
374: %
375: \allowdisplaybreaks{\begin{eqnarray} \mathcal{A}'_{\rm 3PN}&=&
376: \frac{1}{c^6}\left\{-\frac{35\,\dot{r}^6\,\nu}{16} +
377: \frac{175\,\dot{r}^6\,\nu^2}{16} -
378: \frac{175\,\dot{r}^6\,\nu^3}{16}+\frac{15\,\dot{r}^4\,\nu\,v^2}{2}
379: \right.\nonumber\\&&\qquad - \left. \frac{135\,\dot{r}^4\,\nu^2\,v^2}{4}
380: + \frac{255\,\dot{r}^4\,\nu^3\,v^2}{8} -
381: \frac{15\,\dot{r}^2\,\nu\,v^4}{2} + \frac{237\,\dot{r}^2\,\nu^2\,v^4}{8}
382: \right.\nonumber\\ &&\qquad -\left. \frac{45\,\dot{r}^2\,\nu^3\,v^4}{2}
383: + \frac{11\,\nu\,v^6}{4} - \frac{49\,\nu^2\,v^6}{4} + 13\,\nu^3\,v^6
384: \right.\nonumber\\ &&\qquad + \left.\frac{m}{r}\left(
385: 79\,\dot{r}^4\,\nu - \frac{69\,\dot{r}^4\,\nu^2}{2} -
386: 30\,\dot{r}^4\,\nu^3 - 121\,\dot{r}^2\,\nu\,v^2 +
387: 16\,\dot{r}^2\,\nu^2\,v^2 \right.\right.\nonumber\\&&\qquad\qquad\quad~
388: +\left.\left. 20\,\dot{r}^2\,\nu^3\,v^2+\frac{75\,\nu\,v^4}{4} +
389: 8\,\nu^2\,v^4 - 10\,\nu^3\,v^4 \right)\right.\nonumber\\ &&\qquad +
390: \left. \frac{m^2}{r^2}\,\left( \dot{r}^2 +
391: \frac{22717\,\dot{r}^2\,\nu}{168} + \frac{11\,\dot{r}^2\,\nu^2}{8} -
392: 7\,\dot{r}^2\,\nu^3 + \frac{615\,\dot{r}^2\,\nu\,\pi^2}{64}
393: \right.\right.\nonumber\\&&\qquad\qquad\quad~ \left.\left. -
394: \frac{20827\,\nu\,v^2}{840} + \nu^3\,v^2 -
395: \frac{123\,\nu\,\pi^2\,v^2}{64} \right)\right.\nonumber\\&&\qquad
396: +\left.\frac{m^3}{r^3}\left( -16 - \frac{1399\,\nu}{12} -
397: \frac{71\,\nu^2}{2} + \frac{41\,\nu\,{\pi }^2}{16}
398: \right)\right\},\nonumber\\ \mathcal{B}'_{\rm 3PN}&=&
399: \frac{1}{c^6}\left\{-\frac{45\,\dot{r}^5\,\nu}{8} + 15\,\dot{r}^5\,\nu^2
400: + \frac{15\,\dot{r}^5\,\nu^3}{4} + 12\,\dot{r}^3\,\nu\,v^2
401: \right.\nonumber\\&&\qquad -\left. \frac{111\,\dot{r}^3\,\nu^2\,v^2}{4}
402: -12\,\dot{r}^3\,\nu^3\,v^2 -\frac{65\,\dot{r}\,\nu\,v^4}{8} +
403: 19\,\dot{r}\,\nu^2\,v^4 + 6\,\dot{r}\,\nu^3\,v^4
404: \right.\nonumber\\&&\qquad\left. + \frac{m}{r}\left(
405: \frac{329\,\dot{r}^3\,\nu}{6} + \frac{59\,\dot{r}^3\,\nu^2}{2} +
406: 18\,\dot{r}^3\,\nu^3 - 15\,\dot{r}\,\nu\,v^2 - 27\,\dot{r}\,\nu^2\,v^2 -
407: 10\,\dot{r}\,\nu^3\,v^2 \right) \right.\nonumber\\&&\qquad
408: +\left.\frac{m^2}{r^2}\,\left( -4\,\dot{r} -
409: \frac{5849\,\dot{r}\,\nu}{840} + 25\,\dot{r}\,\nu^2 + 8\,\dot{r}\,\nu^3
410: - \frac{123\,\dot{r}\,\nu\,\pi^2}{32}
411: \right)\right\}.\label{ABcoeffprime}
412: \end{eqnarray}}\noindent
413: %
414: These gauge-transformed coefficients are useful because they do not
415: yield the usual complications associated with logarithms. However, they
416: must be handled with care in applications such as~\cite{MW03}, since one
417: must ensure that all other quantities in the problem (energy, angular
418: momentum, gravitational-wave fluxes {\it etc.}) are defined in the same
419: specific harmonic gauge avoiding logarithms. In the following we shall
420: no longer use the coordinate system leading to
421: Eqs.~(\ref{ABcoeffprime}). Notably the expression we shall derive below
422: for the Lagrangian will be valid in the ``standard'' harmonic coordinate
423: system in which the equations of motion are given by~(\ref{eom})
424: with~(\ref{ABcoeff}).
425:
426: \section{Lagrangian and Hamiltonian formulations}\label{secIII}
427: %
428: The Lagrangian for the relative center-of-mass motion is obtained from
429: the 3PN center-of-mass equations of motion (\ref{eom})--(\ref{ABcoeff})
430: in which one ignores the radiation-reaction terms at the 2.5PN and 3.5PN
431: orders. We are indeed interested in the conservative part of the
432: equations of motion, excluding the terms associated with gravitational
433: radiation; only the conservative part is deducible from a Lagrangian. It
434: is known that the Lagrangian in harmonic coordinates will necessarily be
435: a generalized one (from the 2PN order), {\it i.e.} depending not only on
436: the positions and velocities of the particles, but also on their
437: \textit{accelerations}~\cite{DD81b,DS85}. It is also
438: known~\cite{DS85,ABF01} that one can always restrict ourselves to
439: a Lagrangian that is linear in the accelerations.
440:
441: The conservative part of the center-of-mass equations of motion
442: (\ref{eom})--(\ref{ABcoeff}) then take the form (after systematic
443: order-reduction of the accelerations) of the generalized Lagrange
444: equations
445: %
446: \begin{equation}\label{Lequ}
447: \frac{\partial L}{\partial x^i} -\frac{d}{dt}\left(\frac{\partial
448: L}{\partial v^i}\right) +\frac{d^2}{dt^2}\left(\frac{\partial
449: L}{\partial a^i}\right)=\mathcal{O}\left( \frac{1}{c^8} \right)\;,
450: \end{equation}
451: %
452: where $L[x^i, v^i, a^i]$ denotes the generalized center-of-mass
453: Lagrangian --- which is linear in the accelerations appearing at 2PN and
454: 3PN orders. We recall that there is a large freedom for choosing a
455: Lagrangian because we can always add to it the total time derivative of
456: an arbitrary function. As a matter of convenience, we choose below a
457: particular center-of-mass Lagrangian that is ``close'' (in the sense
458: that many coefficients are identical) to some ``fictitious'' Lagrangian
459: obtained from the general-frame one given in Ref.~\cite{ABF01} by the
460: mere Newtonian center-of-mass replacements $y_1^i\rightarrow
461: \frac{m_2}{m} x^i$, $y_2^i\rightarrow -\frac{m_1}{m} x^i$. We point out
462: that such a fictitious Lagrangian is {\it not} the correct Lagrangian
463: for describing the center-of-mass relative motion. Indeed, the actual
464: relations connecting the center-of-mass variables $y_1^i$ and $y_2^i$ to
465: the relative position $x^i$ and velocity $v^i$, involve many
466: post-Newtonian corrections, so the actual center-of-mass Lagrangian must
467: contain some extra terms in addition to those of the latter fictitious
468: one. However, we find that these extra terms arise only from the 2PN
469: order. Our result (when divided by the reduced mass $\mu=m\nu$) is then
470: %
471: \allowdisplaybreaks{\begin{eqnarray}\label{Lharm} \frac{L}{\mu}&=&
472: \frac{v^2}{2} + \frac{m}{r} \nonumber\\ &&
473: +\frac{1}{c^2}\bigg\{\frac{v^4}{8} - \frac{3\,\nu\,v^4}{8} +
474: \frac{m}{r}\,\left( \frac{\dot{r}^2\,\nu}{2} + \frac{3\,v^2}{2} +
475: \frac{\nu\,v^2}{2} \right)-\frac{m^2}{2\,r^2} \bigg\}\nonumber\\ &&
476: +\frac{1}{c^4}\bigg\{ \frac{v^6}{16} - \frac{7\,\nu\,v^6}{16} +
477: \frac{13\,\nu^2\,v^6}{16} \nonumber\\ &&\qquad~ + \frac{m}{r}\,\left(
478: \frac{3\,\dot{r}^4\,\nu^2}{8} - \frac{\dot{r}^2\,a_n\,\nu\,r}{8} +
479: \frac{\dot{r}^2\,\nu\,v^2}{4} - \frac{5\,\dot{r}^2\,\nu^2\,v^2}{4} +
480: \frac{7\,a_n\,\nu\,r\,v^2}{8} \right.\nonumber\\ &&\qquad\qquad\quad~ +
481: \left.\frac{7\,v^4}{8} - \frac{5\,\nu\,v^4}{4} - \frac{9\,\nu^2\,v^4}{8}
482: - \frac{7\,\dot{r}\,\nu\,r\,a_v}{4} \right) \nonumber\\ &&\qquad~
483: +\frac{m^2}{r^2}\,\left( \frac{\dot{r}^2}{2} +
484: \frac{41\,\dot{r}^2\,\nu}{8} + \frac{3\,\dot{r}^2\,\nu^2}{2} +
485: \frac{7\,v^2}{4} - \frac{27\,\nu\,v^2}{8} + \frac{\nu^2\,v^2}{2} \right)
486: \nonumber\\ &&\qquad~ +\frac{m^3}{r^3}\,\left( \frac{1}{2} +
487: \frac{15\,\nu}{4} \right)\bigg\}\nonumber\\ &&
488: +\frac{1}{c^6}\bigg\{\frac{5\,v^8}{128} - \frac{59\,\nu\,v^8}{128} +
489: \frac{119\,\nu^2\,v^8}{64} - \frac{323\,\nu^3\,v^8}{128} \nonumber\\
490: &&\qquad~ + \frac{m}{r}\,\left( \frac{5\,\dot{r}^6\,\nu^3}{16} +
491: \frac{\dot{r}^4\,a_n\,\nu\,r}{16} -
492: \frac{5\,\dot{r}^4\,a_n\,\nu^2\,r}{16} -
493: \frac{3\,\dot{r}^4\,\nu\,v^2}{16} \right.\nonumber\\
494: &&\qquad\qquad\quad~\left.+ \frac{7\,\dot{r}^4\,\nu^2\,v^2}{4} -
495: \frac{33\,\dot{r}^4\,\nu^3\,v^2}{16} -
496: \frac{3\,\dot{r}^2\,a_n\,\nu\,r\,v^2}{16} -
497: \frac{\dot{r}^2\,a_n\,\nu^2\,r\,v^2}{16} \right.\nonumber\\
498: &&\qquad\qquad\quad~\left.+ \frac{5\,\dot{r}^2\,\nu\,v^4}{8} -
499: 3\,\dot{r}^2\,\nu^2\,v^4 +\frac{75\,\dot{r}^2\,\nu^3\,v^4}{16} +
500: \frac{7\,a_n\,\nu\,r\,v^4}{8} \right.\nonumber\\
501: &&\qquad\qquad\quad~\left.- \frac{7\,a_n\,\nu^2\,r\,v^4}{2} +
502: \frac{11\,v^6}{16} - \frac{55\,\nu\,v^6}{16} + \frac{5\,\nu^2\,v^6}{2}
503: \right.\nonumber\\ &&\qquad\qquad\quad~ +\left.
504: \frac{65\,\nu^3\,v^6}{16} + \frac{5\,\dot{r}^3\,\nu\,r\,a_v}{12} -
505: \frac{13\,\dot{r}^3\,\nu^2\,r\,a_v}{8} \right.\nonumber\\
506: &&\qquad\qquad\quad~\left.- \frac{37\,\dot{r}\,\nu\,r\,v^2\,a_v}{8} +
507: \frac{35\,\dot{r}\,\nu^2\,r\,v^2\,a_v}{4} \right) \nonumber\\ &&\qquad~
508: + \frac{m^2}{r^2}\,\left( -\frac{109\,\dot{r}^4\,\nu}{144} -
509: \frac{259\,\dot{r}^4\,\nu^2}{36} + 2\,\dot{r}^4\,\nu^3 -
510: \frac{17\,\dot{r}^2\,a_n\,\nu\,r}{6} \right.\nonumber\\
511: &&\qquad\qquad\quad~ +\left. \frac{97\,\dot{r}^2\,a_n\,\nu^2\,r}{12}
512: +\frac{\dot{r}^2\,v^2}{4} - \frac{41\,\dot{r}^2\,\nu\,v^2}{6} -
513: \frac{2287\,\dot{r}^2\,\nu^2\,v^2}{48} \right.\nonumber\\
514: &&\qquad\qquad\quad~ -\left. \frac{27\,\dot{r}^2\,\nu^3\,v^2}{4} +
515: \frac{203\,a_n\,\nu\,r\,v^2}{12} + \frac{149\,a_n\,\nu^2\,r\,v^2}{6}
516: \right.\nonumber\\ &&\qquad\qquad\quad~ +\left. \frac{45\,v^4}{16} +
517: \frac{53\,\nu\,v^4}{24} + \frac{617\,\nu^2\,v^4}{24} -
518: \frac{9\,\nu^3\,v^4}{4} \right.\nonumber\\ &&\qquad\qquad\quad~ -\left.
519: \frac{235\,\dot{r}\,\nu\,r\,a_v}{24} +
520: \frac{235\,\dot{r}\,\nu^2\,r\,a_v}{6} \right) \nonumber\\ &&\qquad~ +
521: \frac{m^3}{r^3}\,\left( \frac{3\,\dot{r}^2}{2} -
522: \frac{12041\,\dot{r}^2\,\nu}{420} + \frac{37\,\dot{r}^2\,\nu^2}{4} +
523: \frac{7\,\dot{r}^2\,\nu^3}{2} - \frac{123\,\dot{r}^2\,\nu\,{\pi }^2}{64}
524: \right.\nonumber\\ &&\qquad\qquad\quad~ +\left. \frac{5\,v^2}{4} +
525: \frac{387\,\nu\,v^2}{70} - \frac{7\,\nu^2\,v^2}{4} +
526: \frac{\nu^3\,v^2}{2} + \frac{41\,\nu\,{\pi }^2\,v^2}{64}
527: \right.\nonumber\\ &&\qquad\qquad\quad~ \left. + 22\,\dot{r}^2\,\nu\,\ln
528: \Big(\frac{r}{r'_0}\Big) - \frac{22\,\nu\,v^2}{3}\ln
529: \Big(\frac{r}{r'_0}\Big) \right)\nonumber\\ &&\qquad~ +
530: \frac{m^4}{r^4}\,\left( -\frac{3}{8}- \frac{18469\,\nu}{840} +
531: \frac{22\,\nu}{3}\ln \Big(\frac{r}{r'_0}\Big) \right) \bigg\}\;.
532: \end{eqnarray}}\noindent
533: %
534: Witness the acceleration terms present at the 2PN and 3PN orders; our
535: notation is $a_n= {\bf a}\cdot{\bf n}$ and $a_v= {\bf a}\cdot{\bf v}$
536: for the scalar products between $a^i=dv^i/dt$ and the direction $n^i$
537: and velocity $v^i$. We recall here that it is in general forbidden to
538: order-reduce the accelerations in a Lagrangian.
539:
540: We next consider the problem of the Hamiltonian associated with (the
541: conservative part of) the equations of
542: motion~(\ref{eom})--(\ref{ABcoeff}). This problem is not straightforward
543: in harmonic coordinates because of the presence of accelerations at 2PN
544: and 3PN orders in the Lagrangian~(\ref{Lharm}). To proceed, the best is
545: again to change coordinates, and transform the harmonic coordinate
546: system into a new system which avoids the appearance of the
547: accelerations terms, and also, as it will turn out, of the logarithms at
548: the 3PN order. This new coordinate change will thus contain a piece
549: which is identical to the one we used to remove the logarithms of the
550: equations of motion in Section~\ref{secII}. However, this new coordinate
551: system is not harmonic; it was introduced long ago by Arnowitt, Deser \&
552: Misner in their study of the Hamiltonian formulation of general
553: relativity and is called the ADM coordinate system. In Ref.~\cite{ABF01}
554: the ``contact'' transformation between the particles' variables in
555: harmonic coordinates and those in ADM coordinates was determined. By
556: contact transformation we mean the relation between the particles'
557: trajectories $y_A^i(t)$ in one coordinate system, and the corresponding
558: trajectories $Y_A^i(t)$ in another, say $\delta
559: y_A^i(t)=Y_A^i(t)-y_A^i(t)$. Notice that in the contact transformation,
560: the time $t$ is to be viewed as a ``dummy'' variable. The contact
561: transformation is not a coordinate transformation between the spatial
562: vectors in both coordinates, but takes also into account the fact that
563: the time coordinate changes as well; {\it i.e.} $\delta
564: y_A^i=\xi^i(y_A)-\xi^0(y_A)v_A^i/c$, where $\xi^\mu(y_A)$ denotes the
565: four-dimensional change between the coordinates, when evaluated at the
566: position $y_A=[t,y_A^i(t)]$.
567:
568: There is a unique contact transformation such that the 3PN
569: harmonic-coordinates Lagrangian~(\ref{Lharm}) is changed into another
570: Lagrangian whose Legendre transform coincides with the 3PN
571: ADM-coordinates Hamiltonian derived in~\cite{JaraS98}
572: (see~\cite{ABF01,BI03CM} for details). In a first stage this yields
573: the expression for the ADM-coordinates Lagrangian, in which we use
574: names appropriate to the ADM variables $X^i=x^i+\delta x^i$, which
575: means the separation distance $R$, the relative square velocity $V^2$,
576: and the radial velocity $\dot{R}={\bf N}\cdot{\bf V}$. This is an
577: ordinary Lagrangian, depending only on the positions and velocities
578: and without accelerations, $L^{\rm ADM}[X^i, V^i]$, and as we said
579: which is free of logarithms at the 3PN order. Its explicit expression
580: is
581: %
582: \allowdisplaybreaks{\begin{eqnarray}\label{LADM} \frac{L^{\rm ADM}}{\mu}
583: &=&\frac{m}{R}+ \frac{V^2}{2} \nonumber\\ &&
584: +\frac{1}{c^2}\bigg\{\frac{V^4}{8} - \frac{3\,\nu\,V^4}{8} +
585: \frac{m}{R}\,\left( \frac{\nu\,\dot{R}^2}{2} + \frac{3\,V^2}{2} +
586: \frac{\nu\,V^2}{2} \right)-\frac{m^2}{2R^2} \bigg\}\nonumber\\ &&
587: +\frac{1}{c^4}\bigg\{\frac{V^6}{16} - \frac{7\,\nu\,V^6}{16} +
588: \frac{13\,\nu^2\,V^6}{16} \nonumber\\&&\qquad~+ \frac{m}{R}\,\left(
589: \frac{3\,\nu^2\,\dot{R}^4}{8} + \frac{\nu\,\dot{R}^2\,V^2}{2} -
590: \frac{5\,\nu^2\,\dot{R}^2\,V^2}{4} + \frac{7\,V^4}{8} -
591: \frac{3\,\nu\,V^4}{2} - \frac{9\,\nu^2\,V^4}{8}
592: \right)\nonumber\\&&\qquad~ + \frac{m^2}{R^2}\,\left(
593: \frac{3\,\nu\,\dot{R}^2}{2} + \frac{3\,\nu^2\,\dot{R}^2}{2} + 2\,V^2 -
594: \nu\,V^2 + \frac{\nu^2\,V^2}{2} \right) \nonumber\\&&\qquad~
595: +\frac{m^3}{R^3}\,\left( \frac{1}{4} + \frac{3\,\nu}{4}
596: \right)\bigg\}\nonumber\\ && +\frac{1}{c^6}\bigg\{ \frac{5\,V^8}{128} -
597: \frac{59\,\nu\,V^8}{128} + \frac{119\,\nu^2\,V^8}{64} -
598: \frac{323\,\nu^3\,V^8}{128}\nonumber\\&&\qquad~+\frac{m}{R}\,\left(
599: \frac{5\,\nu^3\,\dot{R}^6}{16} + \frac{9\,\nu^2\,\dot{R}^4\,V^2}{16} -
600: \frac{33\,\nu^3\,\dot{R}^4\,V^2}{16} + \frac{\nu\,\dot{R}^2\,V^4}{2} -
601: 3\,\nu^2\,\dot{R}^2\,V^4 \right.\nonumber\\&&\qquad\qquad\quad~\left. +
602: \frac{75\,\nu^3\,\dot{R}^2\,V^4}{16} + \frac{11\,V^6}{16} -
603: \frac{7\,\nu\,V^6}{2} + \frac{59\,\nu^2\,V^6}{16} +
604: \frac{65\,\nu^3\,V^6}{16} \right)\nonumber\\&&\qquad~
605: +\frac{m^2}{R^2}\,\left( - \frac{5\,\nu\,\dot{R}^4}{12} +
606: \frac{17\,\nu^2\,\dot{R}^4}{12} + 2\,\nu^3\,\dot{R}^4 +
607: \frac{39\,\nu\,\dot{R}^2\,V^2}{16} - \frac{29\,\nu^2\,\dot{R}^2\,V^2}{8}
608: \right.\nonumber\\&&\qquad\qquad\quad~\left. -
609: \frac{27\,\nu^3\,\dot{R}^2\,V^2}{4} + \frac{47\,V^4}{16} -
610: \frac{15\,\nu\,V^4}{4} - \frac{25\,\nu^2\,V^4}{16} -
611: \frac{9\,\nu^3\,V^4}{4} \right) \nonumber\\&&\qquad~
612: +\frac{m^3}{R^3}\,\left( \frac{77\,\nu\,\dot{R}^2}{16} +
613: \frac{5\,\nu^2\,\dot{R}^2}{4} + \frac{7\,\nu^3\,\dot{R}^2}{2} +
614: \frac{3\,\nu\,\dot{R}^2\,\pi^2}{64} + \frac{13\,V^2}{8}
615: \right.\nonumber\\&&\qquad\qquad\quad~\left. - \frac{409\,\nu\,V^2}{48}
616: - \frac{5\,\nu^2\,V^2}{8} + \frac{\nu^3\,V^2}{2} - \frac{\nu\,{\pi
617: }^2\,V^2}{64} \right) \nonumber\\&&\qquad~ +\frac{m^4}{R^4}\,\left(
618: -\frac{1}{8} - \frac{109\,\nu}{12} + \frac{21\,\nu\,\pi^2}{32} \right)
619: \bigg\}\;.
620: \end{eqnarray}}\noindent
621: %
622: Next we apply the ordinary Legendre transform to obtain the
623: corresponding Hamiltonian, $H^{\rm ADM}[X^i, P^i]$, which is a function
624: of $X^i$ and the conjugate momentum $P^i = \partial L^{\rm ADM}/\partial
625: V^i$. We find
626: %
627: \allowdisplaybreaks{\begin{eqnarray}\label{HADM} \frac{H^{\rm ADM}}{\mu}
628: &=& \frac{P^2}{2} -\frac{m}{R}\nonumber\\ && +\frac{1}{c^2}\bigg\{-
629: \frac{P^4}{8} + \frac{3\,\nu\,P^4}{8} + \frac{m}{R}\left( -
630: \frac{{P_R}^2\,\nu}{2} - \frac{3\,P^2}{2} - \frac{\nu\,P^2}{2}
631: \right)+\frac{m^2}{2R^2} \bigg\}\nonumber\\ &&
632: +\frac{1}{c^4}\bigg\{\frac{P^6}{16} - \frac{5\,\nu\,P^6}{16} +
633: \frac{5\,\nu^2\,P^6}{16} \nonumber\\&&\qquad~+ \frac{m}{R}\left( -
634: \frac{3\,{P_R}^4\,\nu^2}{8} - \frac{{P_R}^2\,P^2\,\nu^2}{4} +
635: \frac{5\,P^4}{8} - \frac{5\,\nu\,P^4}{2} - \frac{3\,\nu^2\,P^4}{8}
636: \right)\nonumber\\&&\qquad~ + \frac{m^2}{R^2}\,\left(
637: \frac{3\,{P_R}^2\,\nu}{2} + \frac{5\,P^2}{2} + 4\,\nu\,P^2 \right)
638: \nonumber\\&&\qquad~+\frac{m^3}{R^3}\left( -\frac{1}{4} -
639: \frac{3\,\nu}{4} \right) \bigg\}\nonumber\\ &&
640: +\frac{1}{c^6}\bigg\{-\frac{5\,P^8}{128} + \frac{35\,\nu\,P^8}{128} -
641: \frac{35\,\nu^2\,P^8}{64} + \frac{35\,\nu^3\,P^8}{128}
642: \nonumber\\&&\qquad~ + \frac{m}{R}\left( -\frac{5\,{P_R}^6\,\nu^3}{16} +
643: \frac{3\,{P_R}^4\,P^2\,\nu^2}{16} - \frac{3\,{P_R}^4\,P^2\,\nu^3}{16} +
644: \frac{{P_R}^2\,P^4\,\nu^2}{8} \right.\nonumber\\&&\qquad\qquad\quad~
645: \left. - \frac{3\,{P_R}^2\,P^4\,\nu^3}{16}-\frac{7\,P^6}{16} +
646: \frac{21\,\nu\,P^6}{8} - \frac{53\,\nu^2\,P^6}{16} -
647: \frac{5\,\nu^3\,P^6}{16} \right) \nonumber\\&&\qquad~ +
648: \frac{m^2}{R^2}\,\left( \frac{5\,{P_R}^4\,\nu}{12} +
649: \frac{43\,{P_R}^4\,\nu^2}{12} + \frac{17\,{P_R}^2\,P^2\,\nu}{16}
650: \right.\nonumber\\&&\qquad\qquad\quad~ \left.+
651: \frac{15\,{P_R}^2\,P^2\,\nu^2}{8} - \frac{27\,P^4}{16} +
652: \frac{17\,\nu\,P^4}{2} + \frac{109\,\nu^2\,P^4}{16} \right)
653: \nonumber\\&&\qquad~ + \frac{m^3}{R^3}\,\left(
654: -\frac{85\,{P_R}^2\,\nu}{16} - \frac{7\,{P_R}^2\,\nu^2}{4} -
655: \frac{25\,P^2}{8} - \frac{335\,\nu\,P^2}{48}
656: \right.\nonumber\\&&\qquad\qquad\quad~ \left.- \frac{23\,\nu^2\,P^2}{8}
657: - \frac{3\,{P_R}^2\,\nu\,\pi^2}{64} + \frac{\nu\,P^2\,\pi^2}{64}
658: \right)\nonumber\\&&\qquad~ + \frac{m^4}{R^4}\, \left( \frac{1}{8} +
659: \frac{109\,\nu}{12} - \frac{21\,\nu\,\pi^2}{32} \right) \bigg\}\;.
660: \end{eqnarray}}\noindent
661: %
662: We denote $P^2 = {\bf P}^2$ and $P_R = {\bf N}\cdot{\bf P}$. The
663: previous result is in perfect agreement with the center-of-mass
664: Hamiltonian derived in Ref.~\cite{JaraS98}.
665:
666: \section{Dynamical stability of circular orbits}\label{secIV}
667: %
668: As an application let us investigate the problem of the stability,
669: against dynamical gravitational perturbations, of circular orbits at
670: the 3PN order. We want in particular to discuss the existence (or
671: non-existence) of an innermost stable circular orbit (ISCO) at various
672: post-Newtonian orders, which would constitute the analogue for two
673: black holes with comparable masses of the famous orbit
674: $r_\mathrm{ISCO}=6M/c^2$ in the Schwarzschild metric. Notice that we
675: are concerned here with the stability of the orbit with respect to
676: \textit{purely gravitational} perturbations appropriate to the motion
677: of black holes; however it is known that for neutron stars instead of
678: black holes the ISCO is determined by the hydrodynamical instability
679: rather than by the effect of general relativity.
680:
681: We propose to use two different methods for this problem, one based on
682: a perturbation at the level of the equations of motion
683: (\ref{eom})--(\ref{ABcoeff}) in harmonic coordinates, the other one
684: consisting of perturbing the Hamiltonian equations in ADM coordinates
685: for the Hamiltonian~(\ref{HADM}). We shall find a criterion for the
686: stability of orbits and shall present it in the form of an invariant
687: expression (which is the same in different coordinate systems). We
688: shall check that our two methods agree on the result.
689:
690: We deal first with the perturbation of the equations of motion,
691: following the approach proposed in Section III.A of
692: Ref.~\cite{KWW93}. We introduce polar coordinates $(r,\varphi)$ in the
693: orbital plane and pose $u= {\dot r}$ and $\omega= {\dot
694: \varphi}$. Then Eq.~(\ref{eom}) yields the system of equations
695: %
696: \begin{subequations}\label{36}\begin{eqnarray}
697: {\dot r} &=& u\;,\label{36a}\\ {\dot u} &=&
698: -\frac{m}{r^2}\Big[1+\mathcal{A}+\mathcal{B}u\Big]+r\omega^2\;,\label{36b}\\
699: {\dot \omega} &=& -\omega\left[\frac{m}{r^2}\mathcal{B}+\frac{2
700: u}{r}\right]\;,\label{36c}
701: \end{eqnarray}\end{subequations}
702: %
703: where $\mathcal{A}$ and $\mathcal{B}$ are given by Eqs.~(\ref{ABcoeff})
704: as functions of $r$, $u$ and $\omega$ (through $v^2=u^2+r^2\omega^2$).
705: In the case of an orbit which is circular, apart from the adiabatic
706: inspiral due to the 2.5PN and 3.5PN radiation-reaction effects, we have
707: $\dot r=\dot u=\dot \omega=0$ hence $u=0$. Eq.~(\ref{36b}) gives thereby
708: the angular velocity $\omega_0$ of the circular orbit as
709: %
710: \begin{equation}\label{37}
711: \omega_0^2 = \frac{m}{r_0^3}\big(1+\mathcal{A}_0\big)\;.
712: \end{equation}
713: %
714: Solving iteratively this relation at the 3PN order using the equations
715: of motion (\ref{eom})--(\ref{ABcoeff}) we obtain $\omega_0$ as a
716: function of the circular-orbit radius $r_0$ in harmonic coordinates (the
717: result agrees with the one of Refs.~\cite{BF00,BFeom}),
718: %
719: \begin{eqnarray}\label{42}
720: \omega_0^2 =
721: \frac{m}{r_0^3}\Bigg\{1&+&\frac{m}{r_0\,c^2}\Big(-3+\nu\Big)
722: +\frac{m^2}{r_0^2\, c^4}\left(6+\frac{41}{4}\nu+\nu^2\right)\nonumber\\
723: &+&\frac{m^3}{r_0^3\,c^6}\left(-10
724: +\left[-\frac{75707}{840}+\frac{41}{64}\pi^2+22\ln
725: \Big(\frac{r_0}{r'_0}\Big)\right]\nu+\frac{19}{2}\nu^2+\nu^3\right) +
726: \mathcal{O}\left(\frac{1}{c^8}\right)\Bigg\}\;.
727: \end{eqnarray}
728: %
729: The circular-orbit radius $r_0$ should not be confused with the constant
730: $r'_0$ entering the logarithm at the 3PN order and which is issued from
731: Eqs.~(\ref{ABcoeff}).
732:
733: Now we investigate the equations of linear perturbations around the
734: circular orbit defined by the constants $r_0$, $u_0=0$, or, rather, if
735: we were to include the radiation-reaction damping,
736: $u_0=\mathcal{O}(c^{-5})$, and $\omega_0$. We pose
737: %
738: \begin{subequations}\label{38}\begin{eqnarray}
739: r &=& r_0 + \delta r\;,\\
740: u &=& \delta u\;,\\
741: \omega &=& \omega_0 + \delta \omega\;,
742: \end{eqnarray}\end{subequations}
743: %
744: where $\delta r$, $\delta u$ and $\delta \omega$ denote some
745: perturbations of the circular orbit. Then a system of linear equations
746: follows as
747: %
748: \begin{subequations}\label{39}\begin{eqnarray}
749: \dot{\delta r} &=& \delta u\;,\\ \dot{\delta u} &=& \alpha_0\, \delta r
750: + \beta_0\, \delta \omega\;,\\ \dot{\delta \omega} &=& \gamma_0\, \delta
751: u\;,
752: \end{eqnarray}\end{subequations}
753: %
754: where the coefficients, which solely depend on the unperturbed circular
755: orbit, read~\cite{KWW93}
756: %
757: \begin{subequations}\label{40}\begin{eqnarray}
758: \alpha_0 &=& 3 \omega_0^2 - \frac{m}{r_0^2}\left(\frac{\partial
759: \mathcal{A}}{\partial r}\right)_0\;,\\ \beta_0 &=& 2 r_0 \omega_0 -
760: \frac{m}{r_0^2}\left(\frac{\partial \mathcal{A}}{\partial
761: \omega}\right)_0\;,\\ \gamma_0 &=& -\omega_0 \left[\frac{2}{r_0} +
762: \frac{m}{r_0^2}\left(\frac{\partial \mathcal{B}}{\partial
763: u}\right)_0\right]\;.
764: \end{eqnarray}\end{subequations}
765: %
766: In obtaining Eqs. (\ref{40}) we use the fact that $\mathcal{A}$ is a
767: function of the square $u^2$ through $v^2=u^2+r^2\omega^2$, so that
768: $\partial \mathcal{A}/\partial u$ is proportional to $u$ and thus
769: vanishes in the unperturbed configuration (because $u=\delta u$). On the
770: other hand, since the radiation reaction is neglected, $\mathcal{B}$
771: also is proportional to $u$ [see Eqs.~(\ref{ABcoeff})], so only
772: $\partial \mathcal{B}/\partial u$ can contribute at the zeroth
773: perturbative order. Now by examining the fate of perturbations that are
774: proportional to some $e^{i\sigma t}$, we arrive at the condition for the
775: frequency $\sigma$ of the perturbation to be real, and hence for stable
776: circular orbits to exist, as being~\cite{KWW93}
777: %
778: \begin{equation}\label{41}
779: \hat{C}_0 = -\alpha_0 - \beta_0\, \gamma_0 ~> 0\;.
780: \end{equation}
781: %
782: Substituting into this $\mathcal{A}$ and $\mathcal{B}$ at the 3PN order we
783: then arrive at the orbital-stability criterion
784: %
785: \begin{eqnarray}\label{43}
786: \hat{C}_0 = \frac{m}{r_0^3}\Bigg\{1&+&\frac{m}{r_0\,c^2}\Big(-9+\nu\Big)
787: +\frac{m^2}{r_0^2\,c^4}\left(30 +\frac{65}{4}\nu+\nu^2\right)\nonumber\\
788: &+&\frac{m^3}{r_0^3\,c^6}
789: \left(-70+\left[-\frac{29927}{840}-\frac{451}{64}\pi^2+22\ln
790: \Big(\frac{r_0}{r'_0}\Big) \right]\nu+\frac{19}{2}\nu^2+\nu^3\right) +
791: \mathcal{O}\left(\frac{1}{c^8}\right)\Bigg\}\;,
792: \end{eqnarray}
793: %
794: where we recall that $r_0$ is the radius of the orbit in harmonic
795: coordinates.
796:
797: Our second method is to use the Hamiltonian equations based on the 3PN
798: Hamiltonian in ADM coordinates given by Eq.~(\ref{HADM}). We introduce
799: the polar coordinates $(R,\Psi)$ in the orbital plane --- we assume that
800: the orbital plane is equatorial, given by $\Theta=\frac{\pi}{2}$ in the
801: spherical coordinate system $(R,\Theta,\Psi)$ --- and make the
802: substitution
803: %
804: \begin{equation}\label{44}
805: P^2={P_R}^2+\frac{P_\Psi^2}{R^2}\;,
806: \end{equation}
807: %
808: into the Hamiltonian. This yields a ``reduced'' Hamiltonian that is a
809: function of $R$, $P_R$ and $P_\Psi$, namely
810: $\mathcal{H}=\mathcal{H}\big[R,P_R,P_\Psi\big]$, and describes the
811: motion in polar coordinates in the orbital plane (henceforth we denote
812: $\mathcal{H}=H^{\rm ADM}/\mu$). The Hamiltonian equations then read
813: %
814: \begin{subequations}\label{45}\begin{eqnarray}
815: \frac{dR}{dt} &=& \frac{\partial \mathcal{H}}{\partial P_R}\;,\\
816: \frac{d\Psi}{dt} &=& \frac{\partial \mathcal{H}}{\partial P_\Psi}\;,\\
817: \frac{dP_R}{dt} &=& -\frac{\partial \mathcal{H}}{\partial R}\;,\\
818: \frac{dP_\Psi}{dt} &=& 0\;.
819: \end{eqnarray}\end{subequations}
820: %
821: Evidently the constant $P_\Psi$ is nothing but the conserved
822: angular-momentum integral. For circular orbits we have $R=R_0$ (a
823: constant) and $P_R=0$, so
824: %
825: \begin{equation}\label{46}
826: \frac{\partial \mathcal{H}}{\partial R}\big[R_0,0,P_\Psi^0\big] = 0\;,
827: \end{equation}
828: %
829: which gives the angular momentum $P_\Psi^0$ of the circular orbit as a
830: function of $R_0$, and
831: %
832: \begin{equation}\label{47}
833: \omega_0 = \left(\frac{d\Psi}{dt}\right)_0 = \frac{\partial
834: \mathcal{H}}{\partial P_\Psi}\big[R_0,0,P_\Psi^0\big]\;,
835: \end{equation}
836: %
837: which yields the angular frequency of the circular orbit $\omega_0$ ---
838: the same as in Eq. (\ref{42}) --- in terms of $R_0$,
839: %
840: \begin{eqnarray}
841: \omega_0^2 =
842: \frac{m}{R_0^3}\Bigg\{1&+&\frac{m}{R_0\,c^2}\Big(-3+\nu\Big)
843: +\frac{m^2}{R_0^2\,
844: c^4}\left(\frac{21}{4}-\frac{5}{8}\nu+\nu^2\right)\nonumber\\
845: &+&\frac{m^3}{R_0^3\,c^6}\left(-7
846: +\left[-\frac{2015}{48}+\frac{167}{64}\pi^2
847: \right]\nu-\frac{31}{8}\nu^2+\nu^3\right) +
848: \mathcal{O}\left(\frac{1}{c^8}\right)\Bigg\}\;.
849: \end{eqnarray}
850: %
851: The last equation, which is equivalent to $R={\rm const}=R_0$, {\it
852: i.e.}
853: %
854: \begin{equation}
855: \frac{\partial \mathcal{H}}{\partial P_R}\big[R_0,0,P_\Psi^0\big] = 0\;,
856: \end{equation}
857: %
858: is automatically verified because $\mathcal{H}$ is a quadratic function
859: of $P_R$ and hence $\partial \mathcal{H}/\partial P_R$ is zero for
860: circular orbits.
861:
862: We consider now a perturbation of the circular orbit defined by
863: %
864: \begin{subequations}\label{49}\begin{eqnarray}
865: P_R &=& \delta P_R\;,\\ P_\Psi &=& P_\Psi^0 + \delta P_\Psi\;,\\ R &=&
866: R_0 + \delta R\;,\\ \omega &=& \omega_0 + \delta \omega\;.
867: \end{eqnarray}\end{subequations}
868: %
869: It is easy to verify that the Hamiltonian equations~(\ref{45}), when
870: worked out at the linearized order, read as
871: %
872: \begin{subequations}\label{50}\begin{eqnarray}
873: \dot{\delta P_R} &=& -\pi_0\, \delta R - \rho_0\, \delta P_\Psi\;,\\
874: \dot{\delta P_\Psi} &=& 0\;,\\ \dot{\delta R} &=& \sigma_0\, \delta
875: P_R\;,\\ \delta \omega &=& \rho_0\, \delta R + \tau_0\, \delta P_\Psi\;,
876: \end{eqnarray}\end{subequations}
877: %
878: where the coefficients, which depend on the unperturbed orbit, are given
879: by
880: %
881: \begin{subequations}\label{51}\begin{eqnarray}
882: \pi_0&=&\frac{\partial^2 \mathcal{H}}{\partial
883: R^2}\big[R_0,0,P_\Psi^0\big]\;,\\ \rho_0&=&\frac{\partial^2
884: \mathcal{H}}{\partial R\, \partial P_\Psi}\big[R_0,0,P_\Psi^0\big]\;,\\
885: \sigma_0&=&\frac{\partial^2 \mathcal{H}}{\partial
886: {P_R}^2}\big[R_0,0,P_\Psi^0\big]\;,\\ \tau_0&=&\frac{\partial^2
887: \mathcal{H}}{\partial {P_\Psi}^2}\big[R_0,0,P_\Psi^0\big]\;.
888: \end{eqnarray}\end{subequations}
889: %
890: By looking to solutions proportional to some $e^{i\sigma t}$ one obtains
891: some real frequencies, and therefore one finds stable circular orbits,
892: if and only if
893: %
894: \begin{equation}\label{52}
895: \hat{C}_0 = \pi_0\, \sigma_0 ~> 0\;.
896: \end{equation}
897: %
898: Using the Hamiltonian (\ref{HADM}) we readily obtain
899: %
900: \begin{eqnarray}\label{53}
901: \hat{C}_0 = \frac{m}{R_0^3}\Bigg\{1&+&\frac{m}{R_0\,c^2}(-9+\nu)
902: +\frac{m^2}{R_0^2\,c^4}\left(\frac{117}{4}
903: +\frac{43}{8}\nu+\nu^2\right)\nonumber\\
904: &+&\frac{m^3}{R_0^3\,c^6}\left(-61+\left[\frac{4777}{48}
905: -\frac{325}{64}\pi^2 \right]\nu-\frac{31}{8}\nu^2+\nu^3\right) +
906: \mathcal{O}\left(\frac{1}{c^8}\right)\Bigg\}\;.
907: \end{eqnarray}
908: %
909: This result does not look the same as our previous result (\ref{43}),
910: but this is simply due to the fact that it depends on the ADM radial
911: separation $R_0$ instead of the harmonic one $r_0$. Fortunately all the
912: material needed to connect $R_0$ to $r_0$ with the 3PN accuracy is
913: known~\cite{BI03CM}. In the case of circular orbits we readily find
914: %
915: \begin{eqnarray}\label{54}
916: R_0 = r_0\Bigg\{1&+&\frac{m^2}{r_0^2\,
917: c^4}\left(-\frac{1}{4}-\frac{29}{8}\nu\right) +\frac{m^3}{r_0^3\,
918: c^6}\left(\left[\frac{3163}{1680} +\frac{21}{32}\pi^2
919: -\frac{22}{3}\ln\Big(\frac{r_0}{r'_0}\Big)\right]\nu
920: +\frac{3}{8}\nu^2\right) +
921: \mathcal{O}\left(\frac{1}{c^8}\right)\Bigg\}\;.
922: \end{eqnarray}
923: %
924: The difference between $R_0$ and $r_0$ is made out of 2PN and 3PN terms
925: only. Inserting Eq. (\ref{54}) into Eq. (\ref{53}) and re-expanding to
926: 3PN order we find that indeed our basic stability-criterion function
927: $\hat{C}_0$ comes out the same with our two methods.
928:
929: Finally let us give to the function $\hat{C}_0$ an invariant meaning by
930: expressing it with the help of the orbital frequency $\omega_0$ of the
931: circular orbit, or, more conveniently, of the frequency-related
932: parameter
933: %
934: \begin{equation}\label{55}
935: x_0 = \left(\frac{m\,\omega_0}{c^3}\right)^{2/3}\;.
936: \end{equation}
937: %
938: From the inverse of Eq.~(\ref{42}) we readily obtain $r_0$ as a function
939: of $x_0$. This allows us to write the criterion for stability as $C_0 >
940: 0$, where $C_0=\frac{m^2}{c^6\,x_0^3}\hat{C}_0$ admits the
941: gauge-invariant form, which will be the same in all coordinate systems,
942: %
943: \begin{equation}\label{57}
944: C_0 = 1-6\,x_0 + 14\,\nu\,x_0^2 +
945: \left(\left[\frac{397}{2}-\frac{123}{16}\pi^2\right]\nu-14\nu^2\right)\,x_0^3
946: + \mathcal{O}\left(x_0^4\,\right)\;.
947: \end{equation}
948: %
949: This form is more interesting than the coordinate-dependent expressions
950: (\ref{43}) or (\ref{53}), not only because of its invariant form, but
951: also because as we see the 1PN term yields exactly the Schwarzschild
952: result that the innermost stable circular orbit or ISCO of a test
953: particle ({\it i.e.} in the limit $\nu\to 0$) is located at $x_{\rm
954: ISCO}=1/6$. Thus we find that, at the 1PN order, but for {\it any} mass
955: ratio $\nu$,
956: %
957: \begin{equation}\label{58}
958: x_{\rm ISCO}^{\rm 1PN} = \frac{1}{6}\;.
959: \end{equation}
960: %
961: One could have expected that some deviations of the order of $\nu$
962: already occur at the 1PN order, but it turns out that only from the 2PN
963: order does one find the occurence of some non-Schwarzschildian
964: corrections proportional to $\nu$. At the 2PN order we obtain
965: %
966: \begin{equation}
967: x_{\rm ISCO}^{\rm 2PN} =
968: \frac{3}{14\nu}\Bigg(1-\sqrt{1-\frac{14\nu}{9}}~\Bigg)\;.
969: \end{equation}
970: %
971: For equal masses this gives $x_{\rm ISCO}^{\rm 2PN}\simeq 0.187$. Notice
972: also that the effect of the finite mass corrections is to increase the
973: frequency of the ISCO with respect to the Schwarzschild result ({\it
974: i.e.} to make it more inward), and we find $x_{\rm ISCO}^{\rm
975: 2PN}=\frac{1}{6}\left[1+\frac{7}{18}\nu+\mathcal{O}(\nu^2)\right]$.
976: Finally, at the 3PN order and for equal masses $\nu=\frac{1}{4}$, we
977: find that according to our criterion all the circular orbits are stable,
978: and there is no ISCO. More generally, we find that at the 3PN order all
979: orbits are stable when the mass ratio is $\nu > \nu_c$ where $\nu_c
980: \simeq 0.183$.
981:
982: Note that the above stability criterion $C_0$ gives an innermost stable
983: circular orbit, when it exists, that is not necessarily the same as ---
984: and actually differs from --- the innermost circular orbit or ICO, which
985: is defined by the point at which the center-of-mass binding energy of
986: the binary for circular orbits reaches its minimum value~\cite{B02ico}.
987: In this respect the present formalism, which is based on systematic
988: post-Newtonian expansions (without using post-Newtonian resummation
989: techniques like Pad\'e approximants~\cite{DIS98}), differs from some
990: ``Schwarzschild-like'' methods such as the effective-one-body
991: approach~\cite{BuonD98} in which the ICO happens to be also an innermost
992: stable circular orbit or ISCO.
993:
994: As a final comment, let us note that the use of a {\it truncated}
995: post-Newtonian series such as Eq.~(\ref{57}) to determine the ISCO is
996: {\it a priori} meaningful only if we are able to bound the neglected
997: error terms. Furthermore, since we are dealing with a stability
998: criterion, it is not completely clear that the higher-order
999: post-Newtonian correction terms, even if they are numerically small,
1000: will not change qualitatively the response of the orbit to the dynamical
1001: perturbation. This is maybe a problem, and which cannot be answered
1002: rigorously with the present formalism. However, in the regime of the
1003: ISCO (when it exists), we have seen that $x_0$ is rather small,
1004: $x_0\simeq 0.2$ (this is also approximately the value for the ICO
1005: computed in Ref.~\cite{B02ico}), which indicates that the neglected
1006: terms in the truncated series (\ref{57}) should not contribute very
1007: much, because they involve at least a factor $x_0^4\simeq 0.002$. On the
1008: other hand, we pointed out that in the limit $\nu\to 0$ the criterion
1009: $C_0$ gives back the correct {\it exact} result, $x_{\rm ISCO}^{\nu\to
1010: 0} = \frac{1}{6}$. This contrasts with the gauge-dependent power series
1011: (\ref{43}) or (\ref{53}) which give only some approximate results. Based
1012: on these observations, we feel that it is reasonable to expect that the
1013: gauge-invariant stability criterion defined by Eq. (\ref{57}) is
1014: physically meaningful.
1015:
1016:
1017: % etc, etc
1018:
1019: % The Appendices part is started with the command \appendix;
1020: % appendix sections are then done as normal sections
1021: % \appendix
1022:
1023: % \section{}
1024: % \label{}
1025:
1026: % The Acknowledgements are also a un-numbered section
1027:
1028: \section*{Acknowledgements}
1029: This article is based on the work~\cite{BI03CM} which has been done in
1030: collaboration with Bala R. Iyer.
1031:
1032: \bibliography{/home/blanchet/Articles/ListeRef/ListeRef}
1033:
1034: %\begin{thebibliography}{00}
1035: % please try to use the bibitem system -
1036: % the references should be in order of citation in the text
1037:
1038: % \bibitem{label}
1039: % Text of bibliographic item
1040:
1041: %\bibitem{label}
1042: %\end{thebibliography}
1043:
1044: \end{document}
1045: