1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  

2: % 

3: % Gravothermal Catastrophe and Tsallis' Generalized 

4: % Entropy of Self-Gravitating Systems II. 

5: %      ---- Thermodynamic propery in stellar polytrope ----

6: % 

7: %                                  Revised version 

8: %                                              after referee comments       

9: % 

10: %                                  Atsushi Taruya \& Masa-aki Sakagami

11: %

12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  

13: \documentclass[12pt]{elsart}

14: %\documentclass[12pt,doublespacing]{elsart}

15: 

16: % Use the option doublespacing or reviewcopy to obtain double line spacing

17: % \documentclass[doublespacing]{elsart}

18: 

19: % if you use PostScript figures in your article

20: % use the graphics package for simple commands

21: % \usepackage{graphics}

22: % or use the graphicx package for more complicated commands

23: % \usepackage{graphicx}

24: % or use the epsfig package if you prefer to use the old commands

25: % \usepackage{epsfig}

26: 

27: % The amssymb package provides various useful mathematical symbols

28: \usepackage{amssymb}

29: \usepackage{graphics}

30: \usepackage{graphicx}

31: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

32: \newcommand{\xx}{\mbox{\boldmath$x$}}

33: \newcommand{\yy}{\mbox{\boldmath$y$}}

34: \newcommand{\vv}{\mbox{\boldmath$v$}}

35: \newcommand{\pp}{\mbox{\boldmath$p$}}

36: \newcommand{\unitm}{m_{\scriptscriptstyle 0}}

37: \newcommand{\unitl}{l_{\scriptscriptstyle 0}}

38: \newcommand{\unitv}{v_{\scriptscriptstyle 0}}

39: \newcommand{\bftau}{\mbox{\boldmath$\tau$}}

40: \newcommand{\Tphys}{T_{\rm phys}}

41: \newcommand{\Cv}{C_{\rm\scriptscriptstyle V}}

42: \newcommand{\drho}{\delta\rho}

43: \newcommand{\dPhi}{\delta\Phi}

44: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

45: \begin{document}

46: 

47: \begin{frontmatter}

48: 

49: % Title, authors and addresses

50: 

51: % use the thanksref command within \title, \author or \address for footnotes;

52: % use the corauthref command within \author for corresponding author footnotes;

53: % use the ead command for the email address,

54: % and the form \ead[url] for the home page:

55: % \title{Title\thanksref{label1}}

56: % \thanks[label1]{}

57: % \author{Name\corauthref{cor1}\thanksref{label2}}

58: % \ead{email address}

59: % \ead[url]{home page}

60: % \thanks[label2]{}

61: % \corauth[cor1]{}

62: % \address{Address\thanksref{label3}}

63: % \thanks[label3]{}

64: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

65: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

66: \title{\large Gravothermal Catastrophe and Tsallis' Generalized 

67: Entropy of Self-Gravitating Systems II. Thermodynamic properties  

68: of stellar polytrope

69: }

70: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

71: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

72: % use optional labels to link authors explicitly to addresses:

73: % \author[label1,label2]{}

74: % \address[label1]{}

75: % \address[label2]{}

76: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

77: \author[taruya]{Atsushi Taruya}

78: \address[taruya]{Research Center for the Early Universe(RESCEU), 

79: School of Science, University of Tokyo, Tokyo 113-0033, Japan}

80: \ead{ataruya@utap.phys.s.u-tokyo.ac.jp}

81: 

82: % ---------------------------

83: 

84: \author[sakagami]{Masa-aki Sakagami}

85: \address[sakagami]{Department of Fundamental Sciences, FIHS, 

86: Kyoto University, Kyoto 606-8501, Japan}

87: \ead{sakagami@phys.h.kyoto-u.ac.jp}

88: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

89: \begin{abstract}

90: In this paper, we continue to investigate the thermodynamic properties of 

91: stellar self-gravitating system arising from the Tsallis generalized entropy. 

92: In particular, physical interpretation of the thermodynamic instability,  

93: as has been revealed by previous paper(Taruya \& Sakagami, 

94: Physica A 307 (2002) 185), is discussed in detail based 

95: on the framework of non-extensive thermostatistics. Examining the 

96: Clausius relation in a quasi-static experiment, we obtain the 

97: standard result of thermodynamic relation that the 

98: physical temperature of the equilibrium non-extensive system is 

99: identified with the inverse of the Lagrange multiplier,  $\Tphys=1/\beta$. 

100: Using this relation, the specific heat of total system is 

101: computed, and confirm the common feature of self-gravitating system 

102: that the presence of negative specific heat leads to the thermodynamic 

103: instability. In addition to the gravothermal instability discovered 

104: previously, the specific heat shows the curious divergent 

105: behavior at the polytrope index $n>3$, suggesting another type of 

106: thermodynamic instability in the case of the system surrounded by the 

107: thermal bath. Evaluating the second variation of free energy, we check 

108: the condition for onset of 

109: this instability and find that the zero-eigenvalue problem of the 

110: second variation of free energy exactly recovers the marginal 

111: stability condition indicated from the specific heat. Thus, the stellar 

112: polytropic system is consistently characterized by the non-extensive 

113: thermostatistics as a plausible thermal equilibrium state. We also 

114: clarify the non-trivial scaling behavior appeared in specific heat

115: and address the origin of non-extensive nature in stellar 

116: polytrope. 

117: \end{abstract}

118: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

119: \begin{keyword}

120: % keywords here, in the form: keyword \sep keyword

121: non-extensive entropy \sep self-gravitating system 

122: \sep gravothermal instability \sep negative specific heat 

123: \sep stellar polytrope 

124: % PACS codes here, in the form: \PACS code \sep code

125: \PACS 05.20.-y, 05.90.+m, 95.30.Tg

126: \end{keyword} 

127: \end{frontmatter}

128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

129: %

130: %

131: %

132: %

133: %

134: %

135: %

136: % main text

137: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

138: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

139: \section{Introduction}

140: \label{sec: intro}

141: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

142: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

143: %

144: %

145: %

146: %

147: %

148: %%%%%%%%%%%%%%%%%%%%%%%%

149: Due to its complexity and peculiarity, 

150: stellar self-gravitating system has long attracted much attention in 

151: the subject of astronomy and astrophysics, and even statistical physics. 

152: For an isolated stellar system, the dynamical equilibrium is rapidly 

153: attained after a few crossing time and the thermodynamic 

154: description provides useful information in characterizing the late-time 

155: behavior of this system. Even in this simplest situation, however, 

156: the equilibrium state 

157: of self-gravitating system shows various interesting phenomena, 

158: which may offer an opportunity to recast the framework of the 

159: thermodynamics and/or statistical mechanics. 

160: 

161: 

162: %%%%%%%%%%%%%%%%%%%%%%%%

163: In earlier paper, applying the Tsallis' generalized entropy\cite{T1988}, 

164: we have studied the thermodynamic instability of self-gravitating 

165: systems\cite{TS2002}. The self-gravitating stellar system confined in a 

166: spherical cavity of radius, $r_e$, exhibits an instability, so-called 

167: {\it gravothermal catastrophe}, 

168: which has been widely accepted as a fundamental physical process and 

169: plays an important role for the long-term evolution of globular clusters

170: \cite{BT1987,EPI1987,MH1997}. 

171: The presence of this instability has been long known since the 

172: pioneer work by Antonov\cite{Antonov1962} and 

173: Lynden-Bell \& Wood\cite{LW1968}. Historically, 

174: the gravothermal catastrophe has been studied on the basis of the maximum 

175: entropy principle for the phase-space distribution function, with a 

176: particular attention to the Boltzmann-Gibbs entropy

177: \cite{Padmanabhan1989,Padmanabhan1990}. 

178: 

179: 

180: %%%%%%%%%%%%%%%%%%%%%%%%

181: In contrast to  previous work, we have applied the 

182: Tsallis-type generalized entropy to seek 

183: the equilibrium criteria for the first time (for comprehensive review of 

184: Tsallis formalism and its application to other field of physics, see 

185: \cite{T1999,AO2001}). Then, 

186: the distribution function of Vlassov-Poisson system can be reduced 

187: to a stellar polytropic system\cite{PP1993,PP1999}. Evaluating the 

188: second variation of entropy around the 

189: equilibrium state and solving the zero-eigenvalue problem, 

190: the criterion for the onset of gravothermal instability is obtained. 

191: The main results of our previous analysis are summarized as follows:

192: 

193: \vspace*{0.5cm}

194: 

195: \noindent

196: {\bf(i)} Local entropy extremum ceases to exist in cases with polytrope 

197: index $n>5$ for sufficiently larger radius of the wall, 

198: $r_e> \lambda_{\rm crit}\,GM^2/(-E)$, and for highly density contrast, 

199: $\rho_c/\rho_e> D_{\rm crit}$, where $M$ and $E$ denote the total mass 

200: and energy of the system, $\rho_c$ and $\rho_e$ mean the density at 

201: center and edge, respectively. 

202: \\

203: {\bf(ii)} The critical values $\lambda_{\rm crit}$ and $D_{\rm crit}$ depend 

204: on the polytrope index, both of which respectively approach 

205: $0.335$ and $709$ in the limit of $n\to\infty$, 

206: consistent with the well-known result adopting the 

207: Boltzmann-Gibbs entropy.  

208: \\

209: {\bf(iii)} The stability/instability criterion obtained from 

210: the second variation of Tsallis entropy exactly matches with the result 

211: from standard turning-point analysis. 

212: 

213: \vspace*{0.5cm}

214: 

215: 

216: %%%%%%%%%%%%%%%%%%%%%%%%

217: While the successful results suggest that 

218: non-extensive generalization of thermodynamics will offer 

219: various astrophysical applications involving long-range nature of 

220: self-gravitating systems, there still remain some important issues 

221: concerning the physical interpretation of thermodynamic instability. 

222: 

223: 

224: %%%%%%%%%%%%%%%%%%%%%%%%

225: Heuristically, the gravothermal instability is 

226: explained by the presence of negative specific heat as follows. 

227: In a fully relaxed gravitating system with sufficiently larger radius, 

228: negative specific heat arises at the inner part of the system  

229: and we have $C_{\rm V,inner}<0$, 

230: while the specific heat at the outer part remains positive, 

231: $C_{\rm V,outer}>0$, since 

232: one can safely neglect the effect of self-gravity. 

233: In this situation, if a tiny heat flow is momentarily supplied from inner 

234: to outer part,  both the inner and the outer parts get hotter after the 

235: hydrostatic readjustment. Now imagine the case, 

236: $C_{\rm V, outer}>|C_{\rm V, inner}|$.  The outer part has so much thermal 

237: inertia that it cannot heat up as fast as the inner part, and thereby the 

238: temperature difference between inner and outer parts increases. As a 

239: consequence, the heat flow never stops, leading to a catastrophe 

240: temperature growth. 

241: 

242: 

243: %%%%%%%%%%%%%%%%%%%%%%%%

244: While the above thought experiment is naive in a sense 

245: that we artificially divide the system into the inner and the outer part,  

246: the argument turns out to capture an essence of the thermodynamic 

247: instability in cases with the Boltzmann-Gibbs entropy.  

248: Evaluating the specific heat explicitly, Lynden-Bell and Wood\cite{LW1968} 

249: showed that the specific heat of the total system should 

250: be greater than zero at the onset of instability, although the central part 

251: of this system still has the negative specific heat. Therefore, one can 

252: naively expect that the self-gravitating system generally exhibits the 

253: thermodynamic instability associated with the negative specific heat and 

254: this could even hold in the system characterized by the non-extensive entropy. 

255: 

256: 

257: %%%%%%%%%%%%%%%%%%%%%%%%

258: To address this issue, 

259: however, we should remember the following two remarks that have been never 

260: clarified. First note that there exists a subtle point concerning 

261: the concept of temperature in the non-extensive thermodynamics. 

262: Framework of the non-extensive formalism is formally 

263: constructed keeping the standard result of thermodynamic relations

264: \cite{CT1991,PP1997,TMP1998}, 

265: however, the physical temperature, $\Tphys$, might not be simply related 

266: to the usual one, i.e, the inverse of Lagrange multiplier,   

267: as has been criticized recently\cite{MNPP2000,AMPP2001}. 

268: This point is in particular important in evaluating the specific heat. 

269: 

270: 

271: %%%%%%%%%%%%%%%%%%%%%%%%

272: Second, as has been mentioned by the pioneer work of Lynden-Bell \& 

273: Wood\cite{LW1968}, self-gravitating system shows various types of 

274: thermodynamic instability. While our early study deals with the 

275: stellar system confined within an adiabatic wall, one may replace the 

276: adiabatic wall with the thermally conducting wall surrounded by 

277: a heat bath. In this situation, assuming the Boltzmann-Gibbs entropy, 

278: Lynden-Bell \& Wood showed that no equilibrium state exists for 

279: sufficiently low temperature and high-density contrast. Note that 

280: even in this case, the presence of negative specific heat plays an 

281: essential role for the appearance of instability. 

282: 

283: 

284: %%%%%%%%%%%%%%%%%%%%%%%%

285: Keeping the above remarks in mind, in this paper, 

286: we focus on the thermodynamic property of self-gravitating systems 

287: characterized by Tsallis' generalized entropy.  

288: For this purpose, we first investigate the thermodynamic temperature of 

289: the self-gravitating system from the Clausius relation. 

290: To clarify the physical interpretation of thermodynamic instability,  

291: the specific heat is computed and a role of negative 

292: specific is discussed in detail. Then we turn to focus on the 

293: thermodynamic instability 

294: in a system surrounded by the heat bath. The stability/instability 

295: criterion is derived from the second variation of free energy and 

296: a geometrical construction of marginal stability condition is discussed. 

297: 

298: %%%%%%%%%%%%%%%%%%%%%%%%

299: While the problem considered here includes some general issues that 

300: are commonly faced with the application of the non-extensive 

301: thermostatistics, we will tackle this problem based on the {\it old} 

302: Tsallis formalism using the standard statistical average, 

303: which is currently un-common 

304: (e.g., \cite{MNPP2000,AMPP2001}\cite{TMP1998}). The reason why 

305: we do not adopt the {\it standard} Tsallis formalism using the normalized 

306: $q$-expectation values is twofold. As has been mentioned in previous paper, 

307: a naive application of the new formalism apparently shows a problematic 

308: difficulty in our case of the maximum entropy principle 

309: (Sec.\ref{sec: polytrope}), while no such difficulties arise when we apply 

310: the earlier formalism (see ref.\cite{TS2002} in details). Another 

311: reason is that 

312: while the new formalism has been deliberately constructed so as to eliminate 

313: the undesirable divergences in some physical systems \cite{TMP1998}

314: especially with fractal nature, 

315: no serious divergences have appeared in our case. 

316: Precisely speaking, the  physical quantities, e.g. mass and energy,

317: may have divergence for some equilibrium configuration of 

318: the self-gravitating system.

319: In order to remedy this divergence, we confine the system within 

320: a spherical wall, which is a standard prescription 

321: in studies for the self-gravitating system \cite{Antonov1962,LW1968}. 

322: Since even the old Tsallis formalism preserves a consistent framework 

323: that recovers the usual thermodynamic structure, 

324: from a more general view of the non-extensive thermostatistics,  

325: we expect that the present analysis still provides a valuable insight to the 

326: thermodynamic stability of stellar self-gravitating system.   

327: Of course, the analysis using normalized $q$-value must play an 

328: important role in the Tsallis' non-extensive framework 

329: and we plan to extend our analysis to the one with the new 

330: formulation near future. In this sense, present work can be 

331: regarded as a preliminary analysis toward the next step. 

332: This point will be discussed in the last part 

333: of this paper, together with some implications.  

334:  

335: 

336: %%%%%%%%%%%%%%%%%%%%%%%%

337: This paper is organized as follows. in section \ref{sec: polytrope},  

338: we recast the problem that finds the most probable state of equilibrium 

339: stellar distribution adopting the Tsallis entropy. 

340: The main part of this paper is section \ref{sec: non-extensive}, in which 

341: the thermodynamic properties of stellar polytrope 

342: are investigated in detail. After identification of the thermodynamic 

343: temperature, the explicit expression for specific heat is presented and 

344: the marginally stability condition for the thermodynamic instability 

345: is investigated in both the adiabatic and the isothermal cases. 

346: In section \ref{sec: free-energy}, thermodynamic instability in a system 

347: surrounded by a thermal bath is re-considered by means of the free energy 

348: and the marginal stability condition is re-derived from the second 

349: variation of free energy. Furthermore, following the preceding results, 

350: the origin of the non-extensive nature in stellar 

351: polytropic system is discussed in section \ref{sec: discussion}. 

352: Finally, section \ref{sec: summary} is devoted to the summary and discussions.

353: %

354: %

355: %

356: %

357: %

358: %

359: %

360: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

361: \section{Stellar polytrope as an extremum state of 

362: Tsallis entropy} 

363: \label{sec: polytrope}

364: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

365: %

366: %

367: %

368: %

369: %

370: %

371: %

372: %

373: %

374: %

375: %

376: %

377: %%%%%%%%%%

378: In this section, we recast the problem finding the most probable 

379: state of equilibrium stellar system, based on the maximum entropy principle. 

380: In our previous study, the entropy for the phase-space 

381: distribution function has been introduced without recourse to the 

382: correct dimensions. Although this does not alter the stability/instability  

383: criterion for the stellar equilibrium state, for the sake of the 

384: completeness and the later analysis, we repeat the same calculation as 

385: shown in ref.\cite{TS2002}, taking fully account of the correct dimensions. 

386: 

387: 

388: %%%%%%%%%%

389: Suppose a system containing $N$ particles which are confined within 

390: a hard sphere of radius $r_e$. For simplicity, each particle is 

391: assumed to have the same mass $\unitm$ and interacts via Newton gravity. 

392: The problem considered here is to find an equilibrium state 

393: in an adiabatic treatment. That is, we investigate the equilibrium 

394: particle distribution in which the particles elastically bounce from the wall, 

395: keeping the energy $E$ and the total mass $M(=N\unitm)$ constant. 

396: 

397: 

398: %%%%%%%%%%

399: For present purpose, it is better to employ the mean-field treatment 

400: that the correlation between particles is smeared out and the 

401: system can be fully characterized by the one-particle distribution function, 

402: $f(\xx,\vv)$, defined in six-dimensional phase-space $(\xx,\vv)$ 

403: \cite{TS2002,BT1987}\cite{Antonov1962,LW1968,Padmanabhan1989,Padmanabhan1990}. 

404: Let us denote the phase-space element as 

405: $h^3(=l_0^3v_0^3)$ with unit length $\unitl$ and unit velocity $\unitv$. 

406: Since the distribution function $f(\xx,\vv)$ counts the number of particles 

407: in a unit cell of phase-space, using the standard definition of the 

408: statistical average, the energy and the total mass are 

409: respectively expressed as follows:

410: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

411: \begin{eqnarray}

412:   \label{eq: energy}

413:   E\,&=&\, K+U \,\equiv \,\,\unitm

414:   \int\, \left\{\frac{1}{2}\, v^{2}

415:   +\,\frac{1}{2}\,\Phi(\xx)\right\}\,f(\xx,\vv)\,\,d^6 \bftau,

416: \\

417: \nonumber\\

418:   \label{eq: mass}

419:   M\,&=&\,\unitm\,N\, \equiv\,\unitm \int\, f(\xx,\vv)\,\,d^6\bftau \,\,, 

420: \end{eqnarray}

421: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

422: with the quantity $\Phi$ being the gravitational potential: 

423: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

424: \begin{equation}

425:   \label{eq: potential}

426:   \Phi(\xx)=-G\,\unitm\,\int\,\,\frac{f(\xx',\vv')}{|\xx-\xx'|}\,\,

427:   d^6\bftau'.

428: \end{equation}

429: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

430: In the above expressions, the dimensionless integral measure $d^6\bftau$ 

431: is introduced: 

432: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

433: \begin{equation}

434:   \label{eq: phase_measure}

435:     d^{6}\bftau \equiv \frac{d^{3}\xx\,d^{3}\vv}{h^3}\,\,\,\,\,\,

436:     ;\,\, h = l_0\,v_0. 

437: \end{equation}

438: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

439: 

440: 

441: %%%%%%%%%%

442: Owing to the maximum entropy principle, we explore 

443: the most probable state maximizing the entropy. The entropy quoted here 

444: is a quantity defined in the phase-space and it counts the number of 

445: possible particle state. We are specifically concerned with the equilibrium 

446: state for the Tsallis entropy \cite{T1988}: 

447: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

448: \begin{equation}

449: \label{eq: S_q}

450:   S_{q}=-\frac{N}{q-1}\int\, \left[\left(\frac{f}{N}\right)^{q}

451:     -\left(\frac{f}{N}\right)\right]\,\,d^6\bftau.  

452: \end{equation}

453: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

454: Maximizing the entropy $S_q$ under the constraints reduces to the 

455: following mathematical problem using Lagrange multipliers $\alpha$ and 

456: $\beta$: 

457: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

458: \begin{eqnarray}

459: \delta S_q-\alpha\,\delta M-\beta\,\delta E=0,

460: \label{eq: 1st_variation}

461: \end{eqnarray}

462: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

463: which leads to \cite{TS2002,PP1993,PP1999}:

464: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

465: \begin{equation}

466:   \label{eq: df_1}

467:   f(\xx,\vv)\,\, =\,\,A\,\,

468: \left[\Phi_0-\Phi(\xx)-\frac{1}{2}v^2\right]^{1/(q-1)}, 

469: \end{equation}

470: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

471: where the constants $A$ and $\Phi_0$ are respectively given by

472: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

473: \begin{equation}

474:  A=N\,\left\{\left(\frac{q-1}{q}\right)\unitm\beta\right\}^{1/(q-1)}, 

475: \,\,\,\,\,\,\,\,\,\,\,\,

476: \Phi_0= \frac{1-(q-1)\unitm\,\alpha}{(q-1)\unitm\,\beta}.

477: \label{eq: A_Phi0_constants}

478: \end{equation}

479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

480: 

481: 

482: %%%%%%%%%%

483: The one-particle distribution function (\ref{eq: df_1}) is often called 

484: {\it stellar polytrope}, which satisfies the polytropic equation of state 

485: \cite{BT1987}\cite{PP1993}. 

486: The density profile $\rho(r)$ and the isotropic pressure $P(r)$ at the 

487: radius $r=|\xx|$ are respectively given by 

488: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

489: \begin{eqnarray}

490:  \rho(r)\, &\equiv& \,\unitm\,\int\,f(\xx,\vv)\,\,\frac{d^3\vv}{h^3} 

491: \nonumber\\

492: &=& \,\,4\sqrt{2}\pi\,\,B\left(\frac{3}{2},\,\frac{q}{q-1}\right)

493:  \,\,\frac{\unitm\,A}{h^3}\,

494:    \left\{\Phi_0-\Phi(r)\right\}^{1/(q-1)+3/2},

495:    \label{eq: def_rho}

496: \end{eqnarray}

497: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

498: and

499: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

500: \begin{eqnarray}

501:  P(r)\,&\equiv &\,\unitm\,\int\,\frac{1}{3}\,v^2\,f(\xx,\vv)\,\,\frac{d^3\vv}{h^3} 

502: \nonumber\\

503: & =&\,\left(\frac{1}{q-1}+\frac{5}{2}\right)^{-1}\, \rho(r)\,

504: \left\{\Phi_0-\Phi(r)\right\},

505:    \label{eq: def_pressure}

506: \end{eqnarray}

507: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

508: with $B(a,b)$ being the $\beta$ function. 

509: Thus, these two equations lead to the relation 

510: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

511: \begin{equation}

512:  P(r) =K_n\, \rho^{1+1/n}(r), 

513: \label{eq: Eq. of state}

514: \end{equation}

515: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

516: with the polytrope index given by

517: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

518: \begin{equation}

519:  n = \frac{1}{q-1}+\frac{3}{2}.    

520: \label{eq: n_q}

521: \end{equation}

522: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

523: In equation (\ref{eq: Eq. of state}), the dimensional constant $K_n$ is 

524: introduced:  

525: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

526: \begin{equation}

527: K_n \equiv \frac{1}{n+1}

528: \left\{4\sqrt{2}\pi B\left(\frac{3}{2},n-\frac{1}{2}\right) 

529: \,\frac{\unitm\,A}{h^3} \right\}^{-1/n}. 

530: \label{eq: def_Kn}

531: \end{equation}

532: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

533: Note that the above quantity is equivalent to the variable $(n-3/2)T/(n+1)$ 

534: defined in ref.\cite{TS2002}.  

535: 

536: 

537: %%%%%%%%%

538: Once provided the distribution function, the equilibrium 

539: configuration can be completely specified by solving the Poisson equation.   

540: Hereafter, we specifically restrict our attention to the spherically 

541: symmetric configuration for $q>1$(or $n>3/2$), in which 

542: the dynamically stable state is safely attainable and 

543: the thermodynamic arguments turn out to capture the physical 

544: relevance \cite{BT1987}. 

545: 

546: 

547: From the gravitational potential (\ref{eq: potential}), it reads 

548: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

549: \begin{equation}

550: \label{eq: poisson_eq}  

551:  \frac{1}{r^2}\frac{d}{dr}\left(r^2\frac{d\Phi(r)}{dr}\right)=

552:   4\pi G \rho(r).

553: \end{equation}

554: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

555: Combining (\ref{eq: poisson_eq}) with (\ref{eq: def_rho}), we 

556: obtain the ordinary differential equation for $\Phi$. 

557: Equivalently, a set of equations which represent 

558: the hydrostatic equilibrium are derived using 

559: (\ref{eq: def_rho}), (\ref{eq: def_pressure}) and (\ref{eq: poisson_eq}):  

560: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

561: \begin{eqnarray}

562: & \frac{dP(r)}{dr}\,=&\,\,-\frac{Gm(r)}{r^2}\,\rho(r), 

563: \label{eq: hydro_1}

564: \\

565: & \frac{dm(r)}{dr}\,=&\,\,4\pi\rho(r)\,r^2.  

566: \label{eq: hydro_2}

567: \end{eqnarray}

568: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

569: The quantity $m(r)$ denotes the mass evaluated at the radius $r$

570: inside the wall. We then introduce the dimensionless quantities: 

571: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

572: \begin{equation}

573: \label{eq: dimensionless}

574:  \rho=\rho_c\,\left[\theta(\xi)\right]^n,\,\,\,\,\,\,

575: r=\left\{\frac{(n+1)P_c}{4\pi G\rho_c^2}\right\}^{1/2}\,\xi, 

576: \end{equation}

577: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

578: which yields the following ordinary differential equation: 

579: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

580: \begin{equation}

581:  \theta''+\frac{2}{\xi}\theta'+\theta^n=0,

582: \label{eq: Lane-emden_eq}

583: \end{equation}

584: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

585: where prime denotes the derivative with respect to $\xi$. 

586: The quantities $\rho_c$ and $P_c$ in (\ref{eq: dimensionless}) 

587: are the density and the pressure at $r=0$, respectively. 

588: To obtain the physically relevant solution of (\ref{eq: Lane-emden_eq}), 

589: we put the following boundary condition:

590: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

591: \begin{equation}

592:  \theta(0)=1, \,\,\,\,\,\,\,\theta'(0)=0.    

593: \label{eq: boundary}

594: \end{equation}

595: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

596: A family of solutions satisfying (\ref{eq: boundary}) is referred to 

597: as the {\it Emden solution}, which is well-known in the subject of 

598: stellar structure (e.g., see Chap.IV of ref.\cite{Chandra1939}).  

599: 

600: 

601: %%%%%%%%%%%%%

602: Figure \ref{fig: profile} shows the numerical solution of equation 

603: (\ref{eq: Lane-emden_eq}) for various polytrope indices, where 

604: the density profile, $\rho(r)/\rho_c$ is plotted 

605: as a function of dimensionless radius, $\xi$.  

606: Clearly, profiles with index $n<5$ rapidly fall off and they 

607: abruptly terminate at finite radius({\it left-panel}), while 

608: the $n\geq5$ cases infinitely continue to extend over the outer 

609: radius({\it right-panel}). As already mentioned in previous paper, 

610: characteristic feature seen in figure \ref{fig: profile} plays 

611: an essential role for the thermodynamic instability associated with 

612: negative specific heat. 

613: 

614: 

615: %%%%%%%%%%%%%

616: For later analysis, it is convenient to introduce the following 

617: set of variables, referred to as homology invariants 

618: \cite{Chandra1939,KW1990}: 

619: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

620: \begin{eqnarray}

621:  u &\equiv& \frac{d\ln m(r)}{d\ln r}=

622: \frac{4\pi r^3\rho(r)}{m(r)}=-\frac{\xi\theta^n}{\theta'},

623: \label{eq: def_u}

624: \\

625: \nonumber\\

626:  v &\equiv&  - \frac{d\ln P(r)}{d\ln r}=

627: \frac{\rho(r)}{P(r)}\,\,\frac{Gm(r)}{r}

628: =-(n+1)\frac{\xi\theta'}{\theta},  

629: \label{eq: def_v}

630: \end{eqnarray}

631: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

632: which reduce the degree of equation (\ref{eq: Lane-emden_eq}) 

633: from two to one. The derivative of these variables with 

634: respect to $\xi$ becomes

635: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

636: \begin{equation}

637: \frac{du}{d\xi} = \left(3-u-\frac{n}{n+1}\,v \right)\,\frac{u}{\xi},

638: ~~~~~~~~

639: \frac{dv}{d\xi} = \left(-1+u+\frac{1}{n+1}\,v \right)\,\frac{v}{\xi}.

640: \label{eq: d(u,v)/dxi}

641: \end{equation}

642: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

643: Equations (\ref{eq: Lane-emden_eq}) can thus be re-written with 

644: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

645: \begin{equation}

646:  \label{eq: uv_eqn}

647:   \frac{u}{v}\,\frac{dv}{du}=\frac{(n+1)(u-1)+v}{(n+1)(3-u)-nv}.   

648: \end{equation}

649: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

650: The corresponding boundary condition to (\ref{eq: boundary}) 

651: becomes  $(u,v)=(3,0)$. Using these variables, the basic thermodynamic 

652: quantities such as the energy and the entropy are evaluated and the 

653: results are summed up in Appendix A, which are subsequently used  

654: in section \ref{sec: non-extensive}. 

655: %

656: %

657: %

658: %

659: %

660: %

661: %

662: %

663: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

664: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

665: \section{Thermodynamic properties of stellar polytrope} 

666: \label{sec: non-extensive}

667: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

668: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

669: %

670: %

671: %

672: In this section, we address our main issue, i.e, the physical 

673: interpretation of gravothermal instability in stellar polytropes, 

674: based on the framework of non-extensive thermodynamics. 

675: In section \ref{subsec: temperature}, 

676: we first discuss the thermodynamic temperature of stellar polytrope 

677: calculating both the heat and the entropy changes 

678: in a quasi-static treatment. Then we evaluate the specific heat in section 

679: \ref{subsec: specific heat}. The connection between 

680: the absence of extremum entropy state and the presence 

681: of negative specific heat is discussed in detail. Further,  

682: we argue that there appears another type of thermodynamic instability, 

683: which is subsequently analyzed by means of the free energy.  

684: %

685: %

686: %

687: %

688: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

689: \subsection{Thermodynamic temperature from the Clausius relation}

690: \label{subsec: temperature}

691: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

692: %

693: %

694: %

695: %

696: %

697: %%%%%%%%%%%%%%%%%%%%%

698: As has been mentioned in section \ref{sec: intro}, the concept of 

699: temperature is non-trivial in non-extensive thermostatistics. This is 

700: because the standard framework of thermodynamics crucially depends on the 

701: assumption of extensivity of entropy. According to the recent claim, the 

702: definition of physical 

703: temperature $\Tphys$ should be altered depending on the choice of 

704: energy constraint and is related to the inverse of the 

705: Lagrange multiplier, $1/\beta$, with {\it some correction factors} 

706: \cite{MNPP2000,AMPP2001}.   

707: Note, however, that this discussion heavily relies on the extensivity 

708: of the energy as well as the thermodynamic zeroth law. 

709: In our present case, the maximum entropy principle was applied subject 

710: to the constraints $E$ and $M$, adopting the standard definition of mean 

711: values (see eqs.(\ref{eq: energy})(\ref{eq: mass})). As a consequence, 

712: the resultant energy $E$ becomes non-extensive and we cannot apply 

713: the above definition. 

714: 

715: 

716: %%%%%%%%%%%%%%%%%%%%%

717: To address the physical temperature in the present case, 

718: we therefore consider the relation between 

719: the heat transfer and entropy change and seek the most plausible candidate 

720: for thermodynamic temperature. That is, we analyze the 

721: variation of equilibrium configuration under fixing the total mass. 

722: Specifically, we deal with the quasi-static variation along an 

723: equilibrium sequence. 

724: 

725: 

726: %%%%%%%%%%%%%%%%%%%%%

727: Let us first write down the heat change. The thermodynamic 

728: first law states that 

729: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

730: \begin{equation}

731:   \label{eq: d'Q}

732:   d'Q\,\, = \,\, dE \,+\, P_e\,dV,

733: \end{equation}

734: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

735: where the operation $d'$ stands for incomplete differentiation. 

736: The subscript $_e$ denotes a quantity evaluated at the edge.  

737: In the spherically symmetric configuration, the second term in 

738: right-hand side of (\ref{eq: d'Q}) becomes $4\pi r_e^{2} P_e\,dr_e$. 

739: As for the first term,  

740: the energy of the stellar polytropic system within the radius $r_e$, 

741: is computed in Appendix A.1. Introducing the dimensionless parameter 

742: $\lambda$, it is expressed in terms of the homology invariants 

743: as follows: 

744: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

745: \begin{eqnarray}

746:   \lambda &\equiv& -\frac{r_eE}{GM^2}=

747: \,-\frac{1}{n-5}\,

748: \left[\,\frac{3}{2}\left\{1-(n+1)\frac{1}{v_e}\right\}

749:   +(n-2)\frac{u_e}{v_e}\,\right], 

750: \label{eq: lambda}

751: \end{eqnarray}

752: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

753: where the quantity with subscript $_{e}$ represents the one evaluated 

754: at the boundary $r=r_e$. 

755: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

756: Using (\ref{eq: lambda}),  the heat change $d'Q$ is rewritten as follows: 

757: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

758: \begin{eqnarray}

759:   &&  d'Q\,=\,d\left(-\lambda\,\frac{GM^2}{r_e}\right)\,

760: +\,4\pi r_e^2\,P_e\,dr_e,

761: \nonumber \\

762: &&~~~~~~~~~~~~~~~

763: =\,\frac{GM^2}{r_e}\,\,\left\{\left(\lambda+\frac{u_e}{v_e}\right)\,\,

764: \frac{dr_e}{r_e}\,-\,\xi_e\,\frac{d\lambda}{d\xi_e}\,\,

765: \frac{d\xi_e}{\xi_e}\right\}, 

766: \label{eq: d'Q_2}

767: \end{eqnarray}

768: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

769: where the relation $4\pi r_e^4 P_e/(GM^2)=u_e/v_e$ is used in the last 

770: line (see definitions (\ref{eq: def_u})(\ref{eq: def_v})). 

771: In the above expression, derivative of $\lambda$ with respect to $\xi_e$ 

772: can be computed with a help of relation (\ref{eq: d(u,v)/dxi}) (see eq.(33) 

773: of ref.\cite{TS2002}): 

774: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

775: \begin{equation}

776:   \xi_e \frac{d\lambda}{d\xi_e}\, =\,\frac{n-2}{n-5}

777: \,\,\frac{g(u_e,\,v_e)}{2v_e} ,

778: \label{eq: d_lambda/d_xi}

779: \end{equation}

780: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

781: where 

782: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

783: \begin{equation}

784:  g(u,v)=4u^2+2uv-\left\{8+3\left(\frac{n+1}{n-2}\right)\right\}u-

785: \frac{3}{n-2}\,v+3\left(\frac{n+2}{n-2}\right).

786: \label{eq: g(u,v)}

787: \end{equation}

788: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

789: 

790: 

791: 

792: %%%%%%%%%%%%%%%%%%%%%

793: Next focus on the change of the entropy. 

794: From (\ref{Appen_A: entropy_uv}) 

795: in Appendix A.2, the entropy of the extremum state is given by 

796: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

797: \begin{equation}

798:   \label{eq: entropy_homology_re}

799:   S_q = \left(n-\frac{3}{2}\right)

800: \left[\,\frac{1}{n-5}\,\,\frac{\beta GM^2}{r_e}\,

801: \left\{2\frac{u_e}{v_e}-(n+1)\frac{1}{v_e}+1

802:         \right\} +N\right].

803: \end{equation}

804: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

805: Hence, the variation of entropy $dS_q$ under fixing the total mass 

806: can be decomposed into the variation of homology 

807: invariants $(u_e,v_e)$, radius $r_e$ and 

808: Lagrange multiplier $\beta$ as follows:

809: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

810: \begin{eqnarray}

811:   dS_q &=& \frac{n-3/2}{n-5}\,\,\frac{\beta GM^2}{r_e}\,

812: \left[\left(\frac{d\beta}{\beta}-\frac{dr_e}{r_e}\right)\,

813:     \left\{2\frac{u_e}{v_e}-(n+1)\frac{1}{v_e}+1\right\}

814: \right.\nonumber\\

815: &&~~~~~~~~~~~~~~~~~~~~~~~~~ \left.

816: +\left\{2\frac{u_e}{v_e}

817:     \left(\frac{du_e}{u_e}-\frac{dv_e}{v_e}\right)

818:     -\frac{n+1}{v_e}\frac{dv_e}{v_e}\right\}   

819: \right].

820:   \label{eq: dS_q_1}

821: \end{eqnarray}

822: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

823: Among these variations, variation of homology invariants is simply 

824: rewritten with $d\xi_e$, through the relation (\ref{eq: d(u,v)/dxi}). 

825: On the other hand, from the mass conservation, the variation of Lagrange 

826: multiplier, $d\beta$ is related to both the variations of homology 

827: invariants and $dr_e$ as follows. Using the condition of hydrostatic 

828: equilibrium at the edge $r_e$, 

829: one can obtain the following relation (see derivation in Appendix A.3): 

830: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

831: \begin{equation}

832: \label{eq: eta}

833: \eta\equiv

834: \left\{\frac{(GM)^{n}(\unitm\beta)^{n-3/2}}{r_e^{n-3}h^{3}}\right\}^{1/(n-1)}

835: \,\,=\,\,\alpha_n\,\left(u_e\,v_e^n\right)^{1/(n-1)}, 

836: \end{equation}

837: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

838: where the constant $\alpha_n$ is given by 

839: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

840: \begin{equation}

841: \label{eq: kappa_n}

842: \alpha_n \,\,=\,\, \left\{\frac{(n-1/2)^{n-3/2}}{16\sqrt{2}\pi^2\,

843: (n+1)^n\,B(3/2,n-1/2)}\right\}^{1/(n-1)}, 

844: \end{equation}

845: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

846: which asymptotically approaches unity, in the limit $n\to+\infty$. 

847: Keeping the total mass $M$ constant, variation of 

848: (\ref{eq: eta}) yields

849: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

850: \begin{equation}

851: \frac{n-3/2}{n-1}\,\,\frac{d\beta}{\beta}-\frac{n-3}{n-1}\,\,\frac{dr_e}{r_e}

852: =\frac{1}{n-1}\left(\frac{du_e}{u_e}+ n\,\,\frac{dv_e}{v_e}\right). 

853: \end{equation}

854: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

855: We then rewrite it with 

856: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

857: \begin{equation}

858:   \label{eq: beta_re}

859:   \frac{d\beta}{\beta}-\frac{dr_e}{r_e}=\frac{1}{n-3/2}\,\,

860: \left(\,-\frac{3}{2}\frac{dr_e}{r_e}+\,\frac{du_e}{u_e}\,+\,n\,\frac{dv_e}{v_e}

861:  \,\right).

862: \end{equation}

863: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

864: Substituting the relation (\ref{eq: beta_re}) into 

865: equation (\ref{eq: dS_q_1}), the dependence of $d\beta/\beta$ can be 

866: eliminated. 

867: Thus, using the relation (\ref{eq: d(u,v)/dxi}), the final form of the 

868: entropy change is expressed in terms of the variations $d\xi_e$ 

869: and $dr_e$. After some manipulation, we obtain

870: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

871: \begin{eqnarray}

872: &&  dS_q = \frac{\beta GM^2}{r_e}\,\left[\,

873:   -\frac{3/2}{n-5}\left(2\,\frac{u_e}{v_e}-\frac{n+1}{v_e}+1\right)

874:   \frac{dr_e}{r_e} -\frac{n-2}{n-5}\,\frac{1}{2v_e}

875: \right.

876: \nonumber\\

877: &&\times  \left.\left\{

878:   4u_e^2+2u_e v_e -\left(8+3\frac{n+1}{n-2}\right)u_e-

879:   \frac{3}{n-2}\,v_e +3\left(\frac{n+1}{n-2}\right)\right\}

880: \frac{d\xi_e}{\xi_e}

881: \right].

882:   \label{eq: dS_q_2}

883: \end{eqnarray}

884: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

885: 

886: 

887: %%%%%%%%%%%%%%%%%%%%%

888: Now, from the knowledge of the expressions $\lambda$ and 

889: $\xi_e(d\lambda/d\xi_e)$, one can easily show that the above equation 

890: is just identical to 

891: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

892:   \begin{equation}

893:     \label{eq: dS_q_3}

894:     dS_q = \frac{\beta GM^2}{r_e}\,\,\left\{\left(\lambda+\frac{u_e}{v_e}\right)

895:     \frac{dr_e}{r_e}\,-\xi_e\,\frac{d\lambda}{d\xi_e}\,\,

896:     \frac{d\xi_e}{\xi_e}\,\right\}.  

897:   \end{equation}

898: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

899: Therefore, comparison between (\ref{eq: dS_q_3}) and (\ref{eq: d'Q_2}) 

900: immediately leads to the following relation: 

901: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

902: \begin{equation}

903: \label{eq: Clausius_eq} 

904:         dS_q \,\,=\, \beta\,d'Q=\beta\,(dE + P_e dV),  

905: \end{equation}

906: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

907: which exactly coincides with the standard result 

908: of {\it Clausius relation} in a quasi-static process.  

909: 

910: 

911: %%%%%%%%%%%%%%%%%%%%%

912: The relation (\ref{eq: Clausius_eq}) strongly suggests that the 

913: thermodynamic temperature $\Tphys$ is identified with the inverse of 

914: Lagrange multiplier, $\Tphys=1/\beta$. At first glance, 

915: the result seems somewhat trivial, since one can easily expect this 

916: relation from the standard thermodynamic relation, 

917: $\partial S_q/\partial E=\beta$, which generally holds even in the 

918: the non-extensive Tsallis formalism \cite{CT1991,PP1997}. 

919: As advocated by many author, however, the relation 

920: $\partial S_q/\partial E=\beta$ does not simply imply the thermodynamic 

921: temperature $\Tphys=1/\beta$ and it might even contradict with 

922: the thermodynamic temperature defined through the 

923: thermodynamic zeroth low \cite{AMPP2001}. 

924: 

925: 

926: %%%%%%%%%%%%%%%%%%%%%

927: On the other hand, in our case of the self-gravitating system, 

928: the thermodynamic temperature $\Tphys=1/\beta$ is mathematically verified 

929: by the integrable condition of the thermodynamic entropy 

930: through the Clausius relation. 

931: Further, it is remarkably found that the relation $\Tphys=1/\beta$ 

932: holds even in the absence of gravity (the limit $G\to0$) and can be proven 

933: through an alternative route. In Appendix B, as a pedagogical example, we 

934: demonstrate that the relation $\Tphys=1/\beta$ is indeed obtained in the 

935: classical gas model using the Carnot cycle. 

936: %

937: %

938: %

939: %

940: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

941: \subsection{Negative specific heat and thermodynamic instability}

942: \label{subsec: specific heat}

943: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

944: %

945: %

946: %

947: %

948: %

949: %

950: %%%%%%%%%%%%%%%%%%%%%

951: Once obtained the thermodynamic temperature, $\Tphys=1/\beta$, 

952: we are in a position to investigate the thermodynamic instability from 

953: the straightforward calculation of the specific heat. 

954: Let us first discuss the qualitative behavior of the specific heat. 

955: By definition, the specific heat at constant volume is given by 

956: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

957: \begin{equation}

958: \label{eq: C_V_1}

959:         \Cv \equiv

960:         \left(\frac{dE}{d\Tphys}\right)_e 

961: =       -\beta^2\,\left(\frac{dE}{d\beta}\right)_e

962: = - \beta^2\,\,\frac{\displaystyle \left(\frac{dE}{d\xi}\right)_e}

963:         {\displaystyle \left(\frac{d\beta}{d\xi}\right)_e}.  

964: \end{equation}

965: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

966: Recall that the dimensionless parameters $\lambda$ and $\eta$ 

967: are respectively proportional to $-E$ and $\beta^{(n-1)/(n-3/2)}$ (see 

968: eqs.(\ref{eq: lambda})(\ref{eq: eta})). This implies that 

969: for a system of constant mass inside a fixed wall, the 

970: qualitative behavior of (\ref{eq: C_V_1}) can be deduced from the 

971: relation between $\eta$ and $\lambda$. 

972: 

973: 

974: %%%%%%%%%%%%%%%%%%%%%

975: Figure \ref{fig: eta_lambda} depicts the trajectories of the Emden 

976: solutions in the $(\eta,\lambda)$-plane with various polytrope 

977: indices. Each point along the trajectory represents an Emden 

978: solution for different value of the radius $r_e$. 

979: From the boundary condition, all the trajectories start from 

980: $(\eta,\lambda)=(0,-\infty)$, corresponding to the origin $r_e=0$. 

981: As gradually increasing the radius, the trajectories first move to 

982: upper-right direction monotonically, as marked by the arrow. 

983: At this stage, the kinematic energy dominates the potential energy and 

984: the system lies in a kinematically thermal state ($\lambda<0$), 

985: indicating the positive specific heat. 

986: For larger radius, while the curves with index $n\leq3$ abruptly terminate,  

987: the trajectories with $n>3$ suddenly change their direction from upper-right 

988: to upper-left. Moreover,  in the case of $n>5$, the trajectory 

989: progressively changes its direction and it finally spirals around a 

990: fixed point. 

991: 

992: 

993: %%%%%%%%%%%%%%%%%%%%%

994: From these observations,  one can roughly infer the existence of the 

995: two types of the thermodynamic instability as follows. At first inflection 

996: point for $n>3$, the specific heat diverges and the signature of $\Cv$ 

997: becomes indefinite. Beyond this point, the specific heat changes from 

998: positive to negative.  This means that the potential energy conversely 

999: dominates the kinetic energy,  indicating the system 

1000: being {\it gravothermal}. In this case, equilibrium state ceases to exist 

1001: for a system in contact with a heat bath, but does still exist for a system 

1002: surrounded by an adiabatic wall. However, for the polytrope 

1003: index $n>5$, the specific heat of the system turns to increase beyond this 

1004: inflection point and it next reaches at the point $d\lambda/d\eta=0$, i.e, 

1005: $\Cv=0$. This means that while the inner part of the system still keeps the 

1006: specific heat negative, the fraction of the outer normal part grows up as 

1007: increasing $r_e$ and it eventually balances with inner gravothermal part. 

1008: Thus, beyond this critical point, no thermal balance is attainable and the 

1009: system becomes gravothermally unstable. This is true even in the system 

1010: surrounded by an adiabatic wall. 

1011: 

1012: 

1013: %%%%%%%%%%%%%%%%%%%%%

1014: Now, let us write down the explicit expression for the specific heat 

1015: $\Cv$. In equation (\ref{eq: C_V_1}), the variation of $\beta$ and $E$ 

1016: with $\xi_e$ can be respectively rewritten with 

1017: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1018: \begin{equation}

1019: \left(\frac{dE}{d\xi}\right)_e = \,-\,\,\frac{GM^2}{r_e}\,\,

1020:         \frac{d\lambda}{d\xi_e},

1021: \label{eq: dEdxi}

1022: \end{equation}

1023: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1024: and

1025: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1026: \begin{equation}

1027: \left(\frac{d\beta}{d\xi}\right)_e = \,\frac{n-1}{n-3/2}\,

1028:         \frac{\beta}{\eta}\,\,\frac{d\eta}{d\xi_e}. 

1029: \label{eq: detadxi}

1030: \end{equation}

1031: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1032: Here, the variable 

1033: $d\lambda/d\xi_{e}$ has been already given in (\ref{eq: d_lambda/d_xi}). 

1034: As for the derivative of $\eta$ with respect to $\xi_e$, we obtain 

1035: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1036: \begin{equation}

1037:   \label{eq: d_eta/d_xi}

1038: \xi_e\,\,\frac{d\eta}{\xi_e}\,\,=\,\,

1039: \left(u_{e}-\frac{n-3}{n-1}\right)\,\eta.   

1040: \end{equation}

1041: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1042: Then the quantity $\Cv$ becomes 

1043: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1044: \begin{eqnarray}

1045:       \Cv \,\,=\,\,\frac{(n-3/2)(n-2)}{(n-1)(n-5)}\,\,

1046:         \frac{\beta GM^2}{r_e}\,\frac{g(u_e,v_e)}{\displaystyle 

1047:           2v_e\,\left(u_e-\frac{n-3}{n-1}\right)}, 

1048: \nonumber 

1049: \end{eqnarray}

1050: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1051: with the function $g(u_e,v_e)$ given by (\ref{eq: g(u,v)}). 

1052: Notice that the above expression is still redundant, since there remains 

1053: the explicit dependence of the variable $\beta$. Eliminating the variable 

1054: $\beta$ by using the relation (\ref{eq: eta}), one finally obtains

1055: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1056: \begin{equation}

1057:   \label{eq: C_v_2}

1058: \frac{\Cv}{N} = \tilde{\alpha}_n\,\left(\frac{h^2}{GMr_e}\right)^{(3/2)/(n-3/2)}

1059: \frac{g(u_e,v_e)}{\displaystyle 2 \left(u_e-\frac{n-3}{n-1}\right)}\,\,

1060: \left(u_e\,\, v_e^{3/2}\right)^{1/(n-3/2)},  

1061: \end{equation}

1062: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1063: where we introduced the new dimensionless constant $\tilde{\alpha}_n$: 

1064: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1065: \begin{equation}

1066:   \label{eq: tilde_alpha_n}

1067:   \tilde{\alpha}_n \equiv \frac{(n-3/2)(n-2)}{(n-1)(n-5)}\,\,

1068: \alpha_n^{1/(n-3/2)}. 

1069: \end{equation}

1070: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1071: Note that in the limit $n\to+\infty$, equation (\ref{eq: C_v_2}) 

1072: consistently recovers the well-known result of isothermal sphere 

1073: (e.g, eq.(39) of ref.\cite{Chavanis2002}):

1074: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1075: \begin{equation}

1076:   \label{eq: C_v_iso}

1077:   \frac{\Cv}{N}\,\,\,\stackrel{n\to+\infty}{\longrightarrow}\,\,\,

1078:   \frac{4u_e^2 + 2u_e v_e -11 u_e +3}{2(u_e-1)}.

1079: \end{equation}

1080: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1081: Comparing (\ref{eq: C_v_2}) with the isothermal limit, the resultant 

1082: expression contains a residual dimensional parameter 

1083: $h$, as well as the quantities $M$ and $r_e$. While the residual 

1084: dependence can be regarded as a natural consequence of the 

1085: non-extensive generalization of the entropy, it would be helpful 

1086: to understand the origin of this scaling in more simplified manner.  

1087: This will be discussed in section \ref{sec: discussion}. 

1088: 

1089: 

1090: %%%%%%%%%%%%%%%%%%%%%

1091: Apart from the residual factor, the expression of specific heat 

1092: (\ref{eq: C_v_2}) clearly reveals the two types of thermodynamic instability 

1093: seen in Figure \ref{fig: eta_lambda}. The inflection point with the infinite 

1094: specific heat, $\Cv\to\pm\infty$ leads to the condition 

1095: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1096: \begin{equation}

1097:   u_e -  \frac{n-3}{n-1}=0,       

1098:   \label{eq: criterion_1}

1099: \end{equation}

1100: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1101: which immediately yields the conclusion that this is only possible for the 

1102: polytrope index $n>3$, consistent with Figure \ref{fig: eta_lambda}. 

1103: On the other hand, critical point with the vanishing specific heat, 

1104: $\Cv=0$ corresponds to the following condition:  

1105: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1106: \begin{equation}

1107:   g(u_e,\,v_e)=0.  

1108:   \label{eq: criterion_2}

1109: \end{equation}

1110: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1111: This is exactly the same condition as obtained from the second variation 

1112: of entropy (see eq.(33) or (53) in ref.\cite{TS2002}). According to the 

1113: previous analysis, the condition (\ref{eq: criterion_2}) represents the 

1114: marginal stability at which the extremum state of the entropy $S_q$ is 

1115: neither maximum nor minimum. This situation turns out to appear when 

1116: the polytrope index $n>5$. 

1117: 

1118: 

1119: %%%%%%%%%%%%%%%%%%%%%

1120: Therefore, we reach a fully satisfactory conclusion that the 

1121: thermodynamic instability found from the second variation of entropy is 

1122: intimately related to the presence of negative specific heat and the 

1123: stability/instability criterion can be 

1124: exactly recovered from the critical point of the thermal balance, $\Cv=0$, 

1125: which is also consistent with the analysis in the Boltzmann-Gibbs limit, 

1126: $n\to\infty$ \cite{LW1968}. The successful result can be regarded as an 

1127: outcome of the correct definition of $\Tphys$. 

1128: As for the transition point with $\Cv\to\pm\infty$, it clearly  

1129: indicates the thermodynamic instability of a system in contact with a 

1130: thermal bath. In next section, by means of the free energy, we confirm that 

1131: the condition (\ref{eq: criterion_1}) indeed represents the marginal 

1132: stability of the system surrounded by a thermal wall and 

1133: beyond this point the system will be unstable. 

1134: 

1135: 

1136: %%%%%%%%%%%%%%%%%%%%%

1137: In Figure \ref{fig: C_V}, by varying the radius $r_e$, 

1138: the normalized specific heat per particle 

1139: $\Cv^{*}/N$ is plotted as a function of density contrast, 

1140: $\rho_c/\rho_e$ 

1141: around the critical polytrope indices $n=3$({\it upper-panels}) and 

1142: $n=5$({\it middle-panels}). Here, the normalized specific heat $\Cv^{*}$ is defined 

1143: by the specific heat $\Cv$ divided by the redundant factor 

1144: $(h^2/GMr_e)^{(3/2)/(n-3/2)}$. Obviously, the transition point 

1145: $\Cv\to\pm\infty$ appears when $n>3$({\it crosses}), 

1146: while the existence of critical point $\Cv=0$ is allowed for 

1147: higher density contrast of $n>5$ cases({\it arrows}). 

1148: The critical values $D_{\rm crit}\equiv(\rho_c/\rho_e)_{\rm crit}$ 

1149: indicated by arrows exactly coincide

1150: with those obtained from the previous analysis (see Table 1 of 

1151: ref.\cite{TS2002}). Lower-panels of Figure \ref{fig: C_V} show the 

1152: specific heat with large polytrope indices $n=10$ and $30$, together with 

1153: the Boltzmann-Gibbs limit ($n\to+\infty$, labeled by {\it iso}). As increasing 

1154: the polytrope 

1155: index $n$, the critical/transition points tend to shift to the lower density 

1156: contrast, while the successive divergent and zero-crossing points appear 

1157: at the higher density contrast, corresponding to the behavior seen in 

1158: Figure \ref{fig: eta_lambda}. 

1159: %

1160: %

1161: %

1162: %

1163: %

1164: %

1165: %

1166: %

1167: %

1168: %

1169: %

1170: %

1171: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1172: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1173: \section{Thermodynamic instability from the second variation of free energy}

1174: \label{sec: free-energy}

1175: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1176: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1177: %

1178: %

1179: %

1180: %

1181: %

1182: %

1183: %

1184: %%%%%%%%%%%%%%%%%%%%%

1185: Previous section reveals that there exists another type of 

1186: thermodynamic instability in which 

1187: the marginal stability is deduced from the 

1188: condition (\ref{eq: criterion_1}).  

1189: In this section, to check the consistency of the non-extensive 

1190: thermostatistics, we reconsider this issue by means of the Helmholtz 

1191: free energy: 

1192: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1193: \begin{equation}

1194: \label{eq: free_energy} 

1195:         F_q \,=\, E \,-\, \Tphys\,S_q. 

1196: \end{equation}

1197: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1198: Adopting the relation $\Tphys=1/\beta$,  

1199: we re-derive the marginal stability condition (\ref{eq: criterion_1}) 

1200: from the second variation of $F_q$. 

1201: 

1202: 

1203: %%%%%%%%%%%%%%%%%%%%%

1204: Consider a system surrounded by the thermally conducting wall 

1205: in contact with a heat bath. Usually, the stable equilibrium state 

1206: should keep the free energy $F_q$ minimum.  

1207: Thus the presence of thermodynamic instability implies the absence of 

1208: minimum free energy, which can be deduced from the signature of the second 

1209: variation $\delta^2F_q$ around the extremum state of free energy. 

1210: Since the non-extensive formalism still verifies the Legendre 

1211: transform structure leading to the standard result of thermodynamic 

1212: relation\cite{CT1991,PP1997}, the extremum state of the free energy 

1213: exactly coincides with that of the entropy. One thus skips to find the 

1214: extremum state of $F_q$ and proceeds to evaluate the second order variation.  

1215: 

1216: 

1217: %%%%%%%%%%%%%%%%%%%%%

1218: In contrast to the adiabatic treatment, we here deal with the density 

1219: perturbation $\rho\to\rho+\delta\rho$, surrounded by a thermal wall. 

1220: To be specific, we evaluate the second variation 

1221: under keeping the radius $r_e$, the total mass $M$ and the 

1222: temperature $\Tphys$ 

1223: constant. Then the variation of energy up to the second order leads to 

1224: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1225: \begin{eqnarray}

1226: \delta E&=& \,\delta\left[\int\left\{\frac{3}{2}\,P(x)+\frac{1}{2}

1227: \rho(x)\Phi(x)\right\} d^3\xx\,\right],

1228: \nonumber \\

1229: \nonumber \\

1230: &=& \int \left\{\frac{3}{2}\,\delta P + 

1231: \frac{1}{2}(\delta\rho\,\Phi+\rho\,\delta\Phi) +\frac{1}{2}\,\delta\rho\,\delta\Phi\,

1232: \right\} d^3\xx.

1233: \label{eq: delta_E}

1234: \end{eqnarray}

1235: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1236: Similarly, using the expression (\ref{Appen_A: tsallis_entropy}) 

1237: in Appendix A.2, the variation of Tsallis entropy becomes 

1238: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1239: \begin{eqnarray}

1240: \delta S_q &=& \,\delta\,\left[

1241: \left(n-\frac{3}{2}\right)\,\left\{N-\beta\,\int\,\,P(x)\,d^3\xx\,\right\}\,

1242: \right],

1243: \nonumber\\

1244: \nonumber\\

1245: &=& -\,\left(n-\frac{3}{2}\right) \, \beta\,\int\, \delta\,P(x) d^3\xx.

1246: \label{eq: delta_S_q}

1247: \end{eqnarray}

1248: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1249: The above expressions include the variation of pressure $\delta P$,    

1250: which can be expanded with a help of the polytropic equation of state 

1251: (\ref{eq: Eq. of state}):  

1252: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1253: \begin{equation}

1254:   \delta P= \left(1+\frac{1}{n}\right)\,\frac{P}{\rho}\,\delta\rho + 

1255:   \frac{1}{2}\,\left(1+\frac{1}{n}\right)\frac{1}{n}\,\frac{P}{\rho^2}\,

1256:   \left(\delta\rho\right)^2. 

1257: \label{eq: delta_P}

1258: \end{equation}

1259: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1260: Combining the above result with equations (\ref{eq: delta_E}) and 

1261: (\ref{eq: delta_S_q}) and collecting the second order terms only, the second 

1262: variation of free energy becomes

1263: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1264: \begin{eqnarray}

1265: \delta^2 F_q = \delta^2 E - \Tphys \,\delta^2 S_q

1266: = \frac{1}{2}\,\int \left\{\,\frac{n+1}{n}\,\frac{P}{\rho^2}\,

1267:   \left(\delta\rho\right)^2\,+\,\delta\rho\,\delta\Phi\,\right\}\,d^3\xx, 

1268: \label{eq: del^2_F_q}

1269: \end{eqnarray}

1270: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1271: where the relation $\Tphys=1/\beta$ is used in the last line. 

1272: Now, restricting our attention to the spherical symmetric perturbation, 

1273: we introduce the following perturbed quantity (see refs.

1274: \cite{TS2002}\cite{Padmanabhan1989}): 

1275: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1276: \begin{equation}

1277: \label{eq: Q(r)}        

1278: \delta\rho(r)=\frac{1}{4\pi r^2}\,\frac{dQ(r)}{dr}.

1279: \end{equation}

1280: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1281: Then the mass conservation $\delta M=0$ implies the boundary condition 

1282: $Q(0)=Q(r_e)=0$. Substituting (\ref{eq: Q(r)}) into (\ref{eq: del^2_F_q}) and 

1283: repeating the integration by part, one finally reaches the following 

1284: quadratic form: 

1285: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1286: \begin{eqnarray}

1287: \label{eq: d^2_F}       

1288:         \delta^2F_q \,=\, -\frac{1}{2}\,\,\int_0^{r_e}dr\,\,

1289:         Q(r)\left[\,\frac{n+1}{n}\,\frac{d}{dr}\left\{\frac{1}{4\pi\,r^2\,\rho}

1290:         \,\left(\frac{P}{\rho}\right) \frac{d}{dr}\right\}+\frac{G}{r^2}

1291:         \right]Q(r).

1292: \end{eqnarray}

1293: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1294: 

1295: 

1296: %%%%%%%%%%%%%%%%%%%%%

1297: Thus, the problem just reduces to the eigenvalue problem and the 

1298: stability of the system can be deduced from the signature of the eigenvalue.  

1299: More specifically, the onset of instability corresponds to 

1300: the marginally stability condition, $\delta^2F_q=0$, and it is 

1301: sufficient to analyze the zero-eigenvalue equation: 

1302: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1303: \begin{equation}

1304: \label{eq: zero_eigenvalue}

1305:         \hat{L}\,\,Q(r)\equiv

1306:         \left[\frac{d}{dr}\left\{\frac{1}{4\pi\,r^2\,\rho}

1307:         \,\left(\frac{P}{\rho}\right) \frac{d}{dr}\right\}+

1308:         \frac{n}{n+1}\,\frac{G}{r^2}\right]\,Q(r)\,=\,0, 

1309: \end{equation}

1310: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1311: with the boundary condition, $Q(0)=Q(r_e)=0$. 

1312: Equation (\ref{eq: zero_eigenvalue}) has quite similar form to the 

1313: zero-eigenvalue equation found in the adiabatic treatment

1314: (see eq.(46) of ref.\cite{TS2002}).  

1315: Except for the non-local term, one can utilize the previous knowledge 

1316: to solve the equation (\ref{eq: zero_eigenvalue}):

1317: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1318: \begin{equation}

1319: \hat{L}\,(4\pi r^3\,\rho) = \frac{n-3}{n+1}\,\frac{G\,m(r)}{r^2},~~~~~~~ 

1320: \hat{L}\,m(r) = \frac{n-1}{n+1}\,\frac{G\,m(r)}{r^2}. 

1321: \end{equation}

1322: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1323: These two equation leads to the ansatz of the solution:   

1324: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1325: \begin{equation}

1326: \label{eq: analytic_solution}

1327: Q(r) = c\left\{\,\,4\pi\,r^3\,\rho(r) -\frac{n-3}{n-1}\,\,m(r)\right\}. 

1328: \end{equation}

1329: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1330: Here, the variable $c$ is an arbitrary constant. The above equation 

1331: (\ref{eq: analytic_solution}) automatically satisfies the boundary 

1332: condition $Q(0)=0$, while the remaining 

1333: condition $Q(r_e)=0$ puts the following constraint: 

1334: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1335: \begin{equation}

1336: \label{eq: stability_criterion}

1337: Q(r_e)=c\left(4\pi\,r_e^3\,\rho_e-\frac{n-3}{n-1}\,M\right)=

1338: c\left(\,u_e-\frac{n-3}{n-1}\,\right)M=0.     

1339: \end{equation}

1340: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1341: Again, we arrive at the satisfactory result that the solution of 

1342: zero-eigenvalue equation exactly recovers the condition 

1343: (\ref{eq: criterion_1}). 

1344: 

1345: 

1346: %%%%%%%%%%%%%%%%%%%%%

1347: Now, remaining task is to show that the second variation 

1348: $\delta^2 F_q$ becomes negative beyond the transition point of 

1349: $\Cv\to\pm\infty$. One can rewrite the expression (\ref{eq: d^2_F}) with 

1350: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1351: \begin{eqnarray}

1352:   \delta^{2}F_q = \frac{1}{2}\,\left(H-1\right)\,\int_0^{r_e}\,\,

1353:   \frac{GQ^2}{r^2}\,dr, 

1354: \nonumber 

1355: \end{eqnarray}

1356: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1357: with the constant $H$ given by

1358: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1359: \begin{equation}

1360:   H\,\equiv\,\frac{\displaystyle \frac{n+1}{n}\,

1361: \int_0^{r_e}\,\frac{1}{4\pi r^2 \rho}\left(\frac{P}{\rho}\right)

1362: \left(\frac{dQ}{dr}\right)^2\,dr}

1363: {\displaystyle \int_0^{r_e}\,\frac{GQ^2}{r^2}\,dr}. 

1364: \label{eq: def_of_H}

1365: \end{equation}

1366: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1367: That is, the condition $H>1$ implies stable local minimum 

1368: state of free energy, while the inequality $H<1$ represents unstable 

1369: local maximum state. Integrating by part, equation (\ref{eq: def_of_H}) 

1370: can be regarded as an eigenvalue equation with eigenvalue, $H$:  

1371: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1372: \begin{equation}

1373:  -\frac{d}{dr}\left\{\frac{1}{4\pi r^2\rho}\left(\frac{P}{\rho}\right)

1374: \frac{dQ}{dr}\right\}  =H\,\,\frac{n}{n+1}\,\,\frac{GQ}{r^2}. 

1375: \label{eq: eigensystem_H}

1376: \end{equation}

1377: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1378: Obviously, equation (\ref{eq: analytic_solution}) becomes the 

1379: solution of above equation with the minimum eigenvalue, $H_{\rm min}=1$, 

1380: if the condition (\ref{eq: stability_criterion}) is fulfilled.  

1381: In this case, solution (\ref{eq: analytic_solution}) can be regarded as 

1382: the ground state of the eigensystem (\ref{eq: eigensystem_H}), 

1383: since the function (\ref{eq: analytic_solution}) does not possess any 

1384: nodes between $[0,r_e]$. Therefore, for a suitably smaller radius 

1385: $r_e$ or a smaller density contrast $\rho_e/\rho_c$ below the transition point, 

1386: the eigenvalue $H$ should be larger than unity. Conversely, from continuity, 

1387: the condition $H<1$  must be satisfied beyond the critical radius.  

1388: 

1389: 

1390: %%%%%%%%%%%%%%%%%%%%%

1391: Finally, using the $(u,v)$-variables, 

1392: the geometrical meaning of onset of thermodynamic instability 

1393: is briefly discussed in similar manner to the adiabatic case. 

1394: In Figure \ref{fig: uv_crit}, the thick solid lines show the Emden 

1395: trajectories with various polytrope indices in $(u,v)$-plane. 

1396: The thin-solid lines in Figure 

1397: \ref{fig: uv_crit} represents the straight lines, $u-(n-3)/(n-1)=0$. 

1398: Since the equilibrium state only exists along the Emden trajectory, 

1399: the condition (\ref{eq: stability_criterion}) is satisfied 

1400: at the intersection of these two solid lines, which is only 

1401: possible for $n>3$. 

1402: On the other hand, as seen in previous section, the equilibrium system 

1403: surrounded by a thermal wall is characterized by the 

1404: three parameters, $r_e$, $M$ and $\beta$(or $\Tphys$), through the relation 

1405: (\ref{eq: eta}). In other words, the system must lie on the curve: 

1406: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1407: \begin{equation}

1408: v=\left(\frac{\eta}{\alpha_n}\right)^{(n-1)/n}\,u^{-1/n}, 

1409: \label{eq: crit_curve}

1410: \end{equation}

1411: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1412: with some constant value $\eta$. 

1413: We have seen in Figure \ref{fig: eta_lambda} that the constant value 

1414: $\eta$ is bounded from above, $\eta\leq\eta_{\rm crit}$. Thus, the 

1415: critical curve (\ref{eq: crit_curve}) with $\eta=\eta_{\rm crit}$ must 

1416: intersect with both the Emden trajectory and the straight line 

1417: $u-(n-3)/(n-1)=0$ simultaneously. This is clearly shown in Figure 

1418: \ref{fig: uv_crit}, 

1419: where the critical curve is plotted as dashed lines. Since the critical 

1420: curves tangentially intersect with Emden solutions, 

1421: it always satisfies the condition $d\eta/d\xi=0$ at the contact point, 

1422: leading to the condition (\ref{eq: criterion_1}) consistently.

1423: 

1424: 

1425: %%%%%%%%%%%%%%%%%%%%%

1426: Table 1 summarizes the dimensionless quantities $\eta_{\rm crit}$ and 

1427: $D_{\rm crit}\equiv(\rho_{c}/\rho_{e})_{\rm crit}$ evaluated at the contact point. 

1428: As increasing the polytrope index $n$, these values asymptotically approach 

1429: the well-known results of Boltzmann-Gibbs limit, 

1430: $\eta_{\rm crit}\to2.52$ and $D_{\rm crit}\to32.1$.

1431: %

1432: %

1433: %

1434: %

1435: %

1436: %

1437: %

1438: %

1439: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1440: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1441: \section{Origin of non-extensive nature in stellar polytrope}

1442: \label{sec: discussion}

1443: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1444: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1445: %

1446: %

1447: %

1448: %

1449: %

1450: %

1451: %

1452: %

1453: %

1454: %%%%%%%%%%%%%%%%%%%%%%%

1455: As has been mentioned in section \ref{subsec: specific heat}, 

1456: specific heat of the stellar polytropic system explicitly depends 

1457: on the residual dimensional parameter $h$,  

1458: in contrast to the isothermal limit (\ref{eq: C_v_iso}). 

1459: In this section, to contact the physical meaning of the non-extensivity 

1460: in stellar polytrope, we discuss the origin of this residual dependence. 

1461: Indeed, the appearance of the residual factor can be recognized as the 

1462: breakdown of 

1463: both the intensivity of temperature and the extensivity of energy and entropy 

1464: as follows. From equation (\ref{eq: Lane-emden_eq}), 

1465: the asymptotic behavior of the Emden solution becomes  

1466: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1467: \begin{eqnarray}

1468: \theta \sim \xi^{-2/(n-1)},~~~~\rho \sim r^{-2n/(n-1)},

1469: ~~~~(\xi, r \rightarrow \infty)

1470: \nonumber

1471: \end{eqnarray}

1472: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1473: so that the mass within a sphere of radius $r$ is given by 

1474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1475: \begin{equation}

1476: M \sim \rho\, r^3 \propto r_e^{(n-3)/(n-1)}~~.

1477: \label{eq: M-r}

1478: \end{equation}

1479: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1480: Then the energy of a virialized stellar system is roughly estimated as 

1481: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1482: \begin{eqnarray}

1483: E \sim \frac{GM^2}{r_e} \propto r_e^{(n-5)/(n-1)} \propto M^{(n-5)/(n-3)},

1484: \nonumber

1485: %\label{eq: E-M}

1486: \end{eqnarray}

1487: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1488: and the relation (\ref{eq: eta}) tells 

1489: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1490: \begin{eqnarray}

1491: \beta \propto r_e^{-(n-3)/(n-1)/(n-3/2)} \propto M^{-1/(n-3/2)}.

1492: \nonumber

1493: %\label{eq: beta-M}

1494: \end{eqnarray}

1495: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1496: These relations clearly show the breakdown of the intensivity of 

1497: temperature and the extensivity of energy, which lead to 

1498: the scaling of the specific heat per mass: 

1499: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1500: \begin{eqnarray}

1501: \frac{C_{\rm V}}{N} = \frac{1}{M} \frac{dE}{d\Tphys} \sim \frac{\beta E}{M}

1502: \propto  M^{-3(n-2)/(n-3)/(n-3/2)}~~. 

1503: \label{eq: C_v-M}

1504: \end{eqnarray} 

1505: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1506: On the other hand, the dimensionless combination $h^2/(GMr_e)$ 

1507: represents the ratio of a typical scale of the stellar system,  

1508: $GMr \sim (GM/r) r^2 \sim v^2 r^2 $, to that of the reference cell,  

1509: $h = v_0 \, l_0$. This behaves as

1510: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1511: \begin{eqnarray}

1512: \frac{h^2}{GMr_e}\, \propto\, \frac{1}{Mr_e}\, \propto\,  M^{2(n-2)/(n-3)}. 

1513: \label{eq: h2/GMr-M}

1514: \end{eqnarray}  

1515: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1516: Thus, these two equations (\ref{eq: C_v-M}) and (\ref{eq: h2/GMr-M}) lead to the 

1517: scaling relation of (\ref{eq: C_v_2}): 

1518: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1519: \begin{equation}

1520: \frac{C_{\rm V}}{N} \, \sim \,\, 

1521: \left(\frac{h^2}{GMr_e}\right)^{(3/2)/(n-3/2)}.

1522: \label{eq: scaling again}

1523: \end{equation}

1524: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1525: Notice that the Clausius relation (\ref{eq: Clausius_eq}) 

1526: suggests that the entropy per unit mass has the same scaling relation: 

1527: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1528: \begin{eqnarray}

1529: \frac{S_q}{M} \sim \frac{\beta E}{M} \sim \frac{C_{\rm V}}{N},

1530: \nonumber

1531: \end{eqnarray}

1532: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1533: Therefore, resultant dependence (\ref{eq: scaling again}) for the stellar 

1534: polytrope can be a natural outcome of the non-extensivity of the entropy.  

1535: 

1536: 

1537: 

1538: %%%%%%%%%%%%%%%%%%%%%%%

1539: In fact, framework of the thermostatistics generally requires an 

1540: introduction of the scale of the unit cell in order to count the available 

1541: number of states in phase spaces. This is even true in the case of the 

1542: isothermal stellar system($n\to+\infty$ or $q\to1$), but, 

1543: the thermodynamic quantities show somewhat peculiar dependence of the scale 

1544: $h$. A typical example is the entropy: 

1545: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1546: \begin{eqnarray}

1547:    S_{\rm BG} = \frac{M}{m_0} 

1548:         \left\{ 

1549:         \left(2 u_e+v_e-\frac{9}{2} \right) 

1550:         - \ln\left(\frac{u_e v_e^{3/2}}{4\pi}\right)    

1551:         - \frac{3}{2} \ln\left(\frac{h^2}{2\pi GMr_e}\right)

1552:         \right\},

1553: \nonumber

1554: \end{eqnarray}

1555: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1556: where $u_e$ and $v_e$ are the homology invariants for the isothermal system. 

1557: The above equation shows that in the Boltzmann-Gibbs limit, 

1558: $h$-dependence of the entropy can be recognized as a matter of choice of 

1559: an additive constant, so that its derivatives, e.g., specific heat, is 

1560: free from the residual dependence. 

1561: 

1562: 

1563: It should be emphasized that the stellar equilibrium 

1564: system recovers the extensivity in the limit $n\to\infty$ and it behaves as  

1565: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1566: \begin{eqnarray}

1567: E \sim M \sim r,\,\,\, C_{\rm V} \sim M.

1568: \label{eq: iso-case}

1569: \end{eqnarray}

1570: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1571: Also, the temperature becomes intensive in this limit. 

1572: Thus, we readily understand that the 

1573: scaling behavior shown in (\ref{eq: C_v_2}) or 

1574: (\ref{eq: scaling again}) has nothing to do with the long-range nature 

1575: of the gravity. Even in the free polytropic gas model in Appendix B, 

1576: the residual dependence emerges as 

1577: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1578: \begin{eqnarray}

1579: \frac{C_{\rm V}}{N} \sim \,\, 

1580: \left\{\frac{h^2}{(P/\rho) V^{2/3}}\right\}^{(3/2)/(n-3/2)}. 

1581: \nonumber

1582: \end{eqnarray}

1583: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1584: It follows that the explicit dependence of the specific heat 

1585: on the reference cell scale $h$ just originates from the  

1586: the non-extensive nature of Tsallis entropy.  

1587: %

1588: %

1589: %

1590: %

1591: %

1592: %

1593: %

1594: %

1595: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1596: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1597: \section{Summary \& Discussions}

1598: \label{sec: summary}

1599: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1600: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1601: %

1602: %

1603: %

1604: %

1605: %

1606: %

1607: %

1608: %

1609: %

1610: %%%%%%%%%%%%%%%%%%

1611: In this paper, thermodynamic properties of 

1612: the stellar self-gravitating system arising from Tsallis' non-extensive 

1613: entropy have been studied in detail. In particular, physical interpretation 

1614: of the thermodynamic instability previously found from the second 

1615: variation of entropy is discussed in detail within a 

1616: framework of the non-extensive thermostatistics. After briefly reviewing the 

1617: equilibrium state of Tsallis entropy, we first address the issues on  

1618: thermodynamic temperature in the case of equilibrium stellar polytrope. 

1619: Analyzing the heat transfer and the entropy change in a quasi-static 

1620: process, standard form of the Clausius relation is derived,  

1621: irrespective of the non-extensivity of entropy. According to this result, 

1622: we explicitly calculate the specific heat and confirm the presence of 

1623: negative specific heat. The onset of instability found in previous work 

1624: just corresponds to the 

1625: zero-crossing point, $\Cv=0$, supporting the fact that the heuristic 

1626: explanation of gravothermal catastrophe 

1627: holds even in the non-extensive thermostatistics. 

1628: 

1629: 

1630: %%%%%%%%%%%%%%%%%%

1631: Further, the analysis of 

1632: specific heat shows divergent behavior at $n>3$, 

1633: suggesting another type of 

1634: thermodynamic instability, which occurs when the system is surrounded 

1635: by a thermal wall. We then turn to the stability analysis 

1636: by means of the Helmholtz free energy. Similar to the previous early work, 

1637: the stability/instability criterion just reduces to the solution of the 

1638: zero-eigenvalue problem and solving the eigenvalue equation, 

1639: we recover the marginal stability condition derived from the 

1640: divergence of specific heat (\ref{eq: criterion_1}). 

1641: 

1642: %%%%%%%%%%%%%%%%%%

1643: In addition to the thermostatistic treatment, we have also discussed 

1644: the origin of non-extensivity in stellar polytrope. The residual 

1645: dependence of the reference scale $h$ appeared in the specific heat 

1646: (\ref{eq: C_v_2}) naturally arises from the non-extensivity of the entropy 

1647: and the resultant scaling dependence can be simply deduced from 

1648: the asymptotic behavior of the Emden solutions.   

1649: 

1650: 

1651: %%%%%%%%%%%%%%%%%%

1652: The stability analysis using the free energy in section 

1653: \ref{sec: free-energy} is consistent with recent claim by 

1654: Chavanis \cite{Chavanis2001}, who has 

1655: investigated the dynamical instability of polytropic gas sphere. 

1656: According to his early paper \cite{Chavanis2002}, 

1657: the thermodynamic stability of stellar system is intimately 

1658: related to the dynamical stability of gaseous system, which has been 

1659: clearly shown in the case of the isothermal distribution. 

1660: Thus, the correspondence between Chavanis' recent result 

1661: \cite{Chavanis2001} and a part of our present analysis 

1662: can be regarded as a generalization of his early work 

1663: to the polytropic system. Note, however, that starting from the Tsallis 

1664: entropy, we extensively discuss the thermodynamic temperature 

1665: and the specific heat of stellar polytrope. 

1666: Therefore, at least, from the thermodynamic point of view, 

1667: our present analysis provides a valuable insight to the stellar 

1668: equilibrium systems.  

1669: 

1670: 

1671: %%%%%%%%%%%%%%%%%%

1672: A particular interest in the thermodynamic relation is 

1673: the Clausius relation (\ref{eq: Clausius_eq}) that has been 

1674: still preserved in the non-extensive stellar system. This is 

1675: indeed consistent with 

1676: the standard thermodynamic relation $\partial S/\partial E=\beta$,  

1677: if one keeps the volume constant. Note also that the relation 

1678: $\partial S/\partial E=\beta$ is readily obtained from 

1679: the standard Legendre transform structure. While we only dealt with 

1680: a specific case with the non-extensive entropy (\ref{eq: S_q}),  

1681: it is well known 

1682: that the standard Legendre transform structure does generally hold 

1683: independently of the functional form of the entropy \cite{PP1997}.  

1684: Hence, our result in turn suggests that the Clausius relation is also 

1685: valid for any stellar system maximizing the entropic functional more  

1686: general than Tsallis'. 

1687: 

1688: 

1689: %%%%%%%%%%%%%%%%%%

1690: At present, the results shown in this paper seems fully consistent with 

1691:   the general framework of the thermostatics. Apart from the thermodynamic 

1692:   instability, the stellar polytropic system can be a plausible thermodynamic 

1693:   equilibrium state, as well as the isothermal stellar distribution. 

1694:   In the isothermal case, existence of the thermodynamic limit has 

1695:   been discussed by de Vega and S\'anchez \cite{deVegaSanchez}: 

1696: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1697: \begin{eqnarray}

1698: M,~V \rightarrow \infty, ~~~\frac{M}{V^{1/3}} = {\rm fixed},

1699: \nonumber

1700: \end{eqnarray}

1701: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1702:  where $V \sim r^3$ is a volume of the system. Recalling the discussion 

1703:  in section \ref{sec: discussion}, the above condition merely reflects 

1704:  the extensivity of the isothermal system (\ref{eq: iso-case}). Thus, 

1705:  similar argument can hold for the non-extensive system. 

1706:  According to the scaling relation (\ref{eq: M-r}), 

1707:  the existence of the thermodynamic limit in stellar polytrope yields 

1708:  the condition: 

1709: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1710: \begin{eqnarray}

1711:   M,~V \rightarrow \infty, ~~~\frac{M}{V^{(n-3)/(3n-3)}} = {\rm fixed}.

1712:   \nonumber

1713: \end{eqnarray} 

1714: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1715: 

1716: 

1717: %%%%%%%%%%%%%%%%%%

1718: Note, however, that this discussion relies on the 

1719: non-uniqueness of the Boltzmann-Gibbs theory, which can be proven only 

1720: mathematically\cite{AR2000}. Indeed, framework of the thermostatistics 

1721: cannot answer the question whether the stellar polytropic distribution 

1722: is really achieved as a thermodynamic equilibrium. To address this issue, 

1723: we must study the detailed process of the long-term stellar dynamical 

1724: evolution.  In the light of this, the analysis using Fokker-Planck model 

1725: or direct N-body simulation can provide an invaluable insight to the 

1726: non-extensive nature of stellar gravitating systems. This issue is now 

1727: in progress and will be presented elsewhere. 

1728: 

1729: 

1730: %%%%%%%%%%%%%%%%%%

1731: Another remaining issue is the re-examination of the present analysis 

1732: from a view of the 'standard' Tsallis formalism 

1733: using the normalized $q$-expectation values. 

1734: Apart from some technical issues on the treatment of the 

1735: maximum entropy principle, one might naively expect that the consistency 

1736: between the statistical and the thermodynamic analysis should be preserved 

1737: even in the new formalism. However, a rather subtle point would be 

1738: the identification of the thermodynamic temperature. As several author 

1739: stated, the standard Clausius relation should be modified in the new 

1740: Tsallis formalism and the resultant form of the expression apparently 

1741: seems to contradict with the 

1742: thermodynamic temperature defined through the thermodynamic zeroth law 

1743: \cite{AMPP2001}. This point will be in particular important in discussing 

1744: the thermodynamic instability and should be clarified along the line of 

1745: our present treatment. 

1746: 

1747: 

1748: \bigskip

1749: This work is supported in part by the Grant-in-Aid for Scientific Research 

1750: of Japan Society for the Promotion of Science. 

1751: %

1752: %

1753: %

1754: %

1755: %

1756: %

1757: %

1758: %

1759: %

1760: %

1761: %

1762: %

1763: %

1764: \clearpage

1765: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1766: %%%%%%%%%%%%%%%%%%%%%%%%%%   Appendix   %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1767: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1768: %

1769: %

1770: %

1771: %

1772: %

1773: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1774: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1775: \section*{Appendix A:  Thermodynamic variables in a stellar polytropic system}

1776: \label{appen_A}

1777: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1778: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1779: %

1780: %

1781: %

1782: In this appendix, using the equilibrium state of stellar polytrope 

1783: described in section \ref{sec: polytrope}, we explicitly evaluate the 

1784: thermodynamic variables,  which have been used in section 

1785: \ref{sec: non-extensive} and \ref{sec: free-energy}.

1786: %

1787: %

1788: %

1789: %

1790: %

1791: % ////////////////////////////////////////////////////////////// %

1792: % ////////////////////////////////////////////////////////////// %

1793: \subsection*{A.1~ Energy}

1794: % ////////////////////////////////////////////////////////////// %

1795: % ////////////////////////////////////////////////////////////// %

1796: %

1797: %

1798: %

1799: %

1800: %

1801: %

1802: Recall that the equilibrium system confined in a spherical container 

1803: satisfies the following virial theorem (e.g, p.502 of Ref.\cite{BT1987}): 

1804: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1805: \begin{eqnarray}

1806:   2K+U = 4\pi r_e^3\,P_e. 

1807: \nonumber

1808: \end{eqnarray}

1809: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1810: The energy (\ref{eq: energy}) is then expressed as 

1811: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1812: \begin{eqnarray}

1813:   E = K+U = 4\pi r_e^3\,P_e - \,K 

1814:     = 4\pi r_e^3\,P_e - \,\frac{3}{2}\int_{0}^{r_e}\,P(r)4\pi\,r^2 dr.

1815: \end{eqnarray}

1816: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1817: To evaluate the above integral in the 

1818: spherically symmetric case, we use the following 

1819: integral formula:  

1820: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1821: \begin{equation}

1822: \int_0^{r_e} P(r)\,4\pi r^2 dr =  -\frac{1}{n-5}\,

1823: \left\{8\pi\,r_e^3\,P_e-(n+1)\,\frac{MP_e}{\rho_e}+

1824: \frac{GM^2}{r_e}\right\}, 

1825: \label{Appen_A: integral}

1826: \end{equation}

1827: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1828: which can be derived from the conditions of hydrostatic equilibrium,  

1829: (\ref{eq: hydro_1}) and (\ref{eq: hydro_2}) 

1830: (see Appendix A of ref.\cite{TS2002}). 

1831: Thus, the energy of extremum state becomes 

1832: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1833: \begin{eqnarray}

1834: E  = \frac{1}{n-5}\,\left[\frac{3}{2} \left\{ \frac{GM^2}{r_e}-

1835:         (n+1)\frac{MP_e}{\rho_e} \right\}+(n-2)\,4\pi r_e^3P_e\right].

1836: \label{Appen_A: energy}

1837: \end{eqnarray}

1838: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1839: In terms of the homology invariants, we obtain 

1840: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1841: \begin{equation}

1842:   E =\frac{1}{n-5}\,\frac{GM^2}{r_e}\,

1843: \left[\,\frac{3}{2}\left\{1-(n+1)\frac{1}{v_e}\right\}

1844:   +(n-2)\frac{u_e}{v_e}\,\right].

1845: \label{Appen_A: energy_uv}

1846: \end{equation}

1847: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1848: %

1849: %

1850: %

1851: %

1852: % ////////////////////////////////////////////////////////////// %

1853: % ////////////////////////////////////////////////////////////// %

1854: \subsection*{A.2~ Entropy}

1855: % ////////////////////////////////////////////////////////////// %

1856: % ////////////////////////////////////////////////////////////// %

1857: %

1858: %

1859: %

1860: %

1861: First note the definition of Tsallis entropy (\ref{eq: S_q}): 

1862: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1863: \begin{eqnarray}

1864:   \nonumber

1865:   S_q = -\left(n-\frac{3}{2}\right)\left\{\int\,

1866: N\,\left(\frac{f}{N}\right)^{(n-1/2)/(n-3/2)}d^6\bftau 

1867: - N \right\}.

1868: \end{eqnarray}

1869: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1870: Substituting the distribution function (\ref{eq: df_1}) into the above 

1871: equation, after some manipulation, we obtain

1872: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1873: \begin{eqnarray}

1874:     S_q  = -\left(n-\frac{3}{2}\right) 

1875:         \,\left\{\beta\int \,P(x)\,d^3\xx -N \right\}.

1876: \label{Appen_A: tsallis_entropy}

1877: \end{eqnarray}

1878: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1879: Thus, the substitution of integral formula (\ref{Appen_A: integral}) 

1880: immediately leads to 

1881: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1882: \begin{eqnarray}

1883: S_q  = \left(n-\frac{3}{2}\right)\left[\frac{1}{n-5}\,

1884:   \left\{8\pi r_e^3 P_e-(n+1)\frac{MP_e}{\rho_e}

1885:         +\frac{GM^2}{r_e}\right\}\beta +N\right],

1886: \nonumber

1887: \end{eqnarray}

1888: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1889: which can be expressed in terms of the homology invariants: 

1890: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1891: \begin{eqnarray}

1892:  \label{Appen_A: entropy_uv}

1893:  S_q = \left(n-\frac{3}{2}\right)

1894:  \left[\,\frac{1}{n-5}\,\,\frac{\beta GM^2}{r_e}\,

1895:    \left\{2\frac{u_e}{v_e}-(n+1)\frac{1}{v_e}+1 \right\} +N\right].

1896: \end{eqnarray}

1897: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1898: %

1899: %

1900: %

1901: % ////////////////////////////////////////////////////////////// %

1902: % ////////////////////////////////////////////////////////////// %

1903: \subsection*{A.3~ Radius-mass-temperature relation }

1904: % ////////////////////////////////////////////////////////////// %

1905: % ////////////////////////////////////////////////////////////// %

1906: %

1907: %

1908: %

1909: The mass-radius-temperature relation (\ref{eq: eta}) is 

1910: derived from the equilibrium stellar polytropic configuration. 

1911: Using (\ref{eq: hydro_1}), we first write down 

1912: the condition of hydrostatic equilibrium at the boundary $r_e$: 

1913: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1914: \begin{eqnarray}

1915:   \frac{GM}{r_{e}^{2}}=-\frac{1}{\rho_e}\,\left(\frac{dP}{dr}\right)_e.

1916: \nonumber

1917: \end{eqnarray}

1918: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1919: The right-hand-side of this  equation is rewritten with 

1920: dimensionless quantities in (\ref{eq: dimensionless}): 

1921: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1922: \begin{equation}

1923:   \label{Appen_A: hydro2}

1924:   \frac{GM}{r_{e}^{2}}=-(n+1)K_{n}\,\rho_c^{1/n}\,\,

1925:   \left(\frac{\xi_e}{r_e}\right)\,\,\theta'_e.

1926: \end{equation}

1927: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1928: We wish to express the above equation only in terms of the variables 

1929: at the edge. To do this, we eliminate the residual dependences, 

1930: $\rho_c$ and $K_n$ from (\ref{Appen_A: hydro2}). The definition 

1931: (\ref{eq: dimensionless}) leads to 

1932: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1933: \begin{eqnarray}

1934: \frac{\xi_e}{r_e}=\left\{ \frac{4\pi G \rho_c^2}{(n+1)P_c}\right\}^{1/2}

1935: =\left\{ \frac{4\pi G}{(n+1)K_n} \right\}^{1/2}\,\,\rho_c^{(n-1)/(2n)}, 

1936: \nonumber

1937: \end{eqnarray}

1938: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1939: which can be rewritten with 

1940: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1941: \begin{eqnarray}

1942: \rho_c^{1/n} = \left\{\frac{4\pi G}{(n+1)K_n}\right\}^{1/(n-1)}\,\,

1943:         \left(\frac{\xi_e}{r_e}\right)^{2/(n-1)}.  

1944: \nonumber

1945: \end{eqnarray}

1946: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1947: Substituting the above relation into (\ref{Appen_A: hydro2}), the 

1948: $\rho_c$-dependence is first eliminated and we obtain

1949: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1950: \begin{equation}

1951: \frac{G^{n/(n-1)}M}{r_e^{(n-3)/(n-1)}}=

1952:         -\left[\frac{\left\{ (n+1)K_n \right\}^n}{4\pi}\right]^{1/(n-1)}\,\,

1953:         \xi_e^{(n+1)/(n-1)}\,\theta'_e.

1954: \label{Appen_A: gmr_relation}

1955: \end{equation}

1956: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1957: As for $K_n$-dependence, the definition (\ref{eq: def_Kn}) together with 

1958: (\ref{eq: A_Phi0_constants}) yields 

1959: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1960: \begin{eqnarray}

1961: (n+1)\,K_n = 

1962:         \left\{ 4\sqrt{2}\pi\,\,

1963: \frac{B(3/2,n-1/2)}{(n-1)^{n-3/2}}\,\,

1964: \frac{M}{h^3}\right\}^{-1/n}(m_0\beta)^{-(n-3/2)/n}.

1965: \label{Appen_A: Kn-beta}

1966: %\nonumber

1967: \end{eqnarray}

1968: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1969: Hence, substituting the above expression into (\ref{Appen_A: gmr_relation}),  

1970: the relation between mass $M$, radius $r_e$ and Lagrange multiplier $\beta$ 

1971: can be finally obtained. In terms of the homology invariants, it follows that  

1972: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1973: \begin{equation}

1974: \left\{\frac{(GM)^{n}(m_0\beta)^{n-3/2}}{r_e^{n-3}h^{3}}\right\}^{1/(n-1)}

1975: \,\,=\,\,\alpha_n\,\left(u_e\,\,v_e^n\right)^{1/(n-1)}, 

1976: \end{equation}

1977: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1978: where the constant $\alpha_n$ is given by

1979: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1980: \begin{eqnarray}

1981: \alpha_n \,\,\equiv\,\, \left\{\frac{(n-1/2)^{n-3/2}}{16\sqrt{2}\pi^2\,

1982: (n+1)^n\,B(3/2,n-1/2)}\right\}^{1/(n-1)},  

1983: \nonumber

1984: \end{eqnarray}

1985: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1986: which asymptotically approaches unity in the limit $n\to\infty$. 

1987: %

1988: %

1989: %

1990: %

1991: %

1992: %

1993: %

1994: %

1995: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1996: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1997: \section*{Appendix B: Thermodynamic temperature of classical gas 

1998:   model from the Carnot cycle}

1999: \label{appen_B}

2000: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2001: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2002: %

2003: %

2004: %

2005: %

2006: %

2007: %

2008: %

2009: %

2010: In a standard framework of thermodynamics, the temperature is defined 

2011: by means of an efficiency of the Carnot cycle. Here we apply the standard 

2012: procedure to seek the physical temperature $T_{\rm phys}$ for so-called 

2013: polytropic system of which distribution function is given by the extremization

2014: of the Tsallis entropy (see eqs.(\ref{eq: S_q})(\ref{eq: 1st_variation})).

2015: For simplicity, we discuss a case of the free classical gas 

2016: without gravity, which corresponds to the $G \rightarrow 0$ limit of 

2017: the {\it stellar polytropic system}. 

2018: 

2019: 

2020: From the $G \rightarrow 0$ limit of the formula (\ref{Appen_A: energy}), 

2021: free polytropic system of the volume $V$ with {\it homogeneous} 

2022: pressure $P$ and density $\rho$ has an (internal) energy: 

2023: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2024: \begin{equation}

2025: E = K = \frac{3}{2}\, P V  = \frac{3}{2}\,\frac{MP}{\rho}. 

2026: \label{Appn_B: energy}

2027: \end{equation}

2028: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2029: Here we drop the subscript ${_e}$ for the pressure and density, since both are 

2030: constant within the system in absence of gravity. 

2031: And equation of state (\ref{eq: Eq. of state}) becomes  

2032: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2033: \begin{equation}

2034: P = K_n ~\rho^{1+1/n}~ = K_n \, \biggl(\frac{M}{V}\biggr)^{1+1/n}.

2035: \label{Appen_B: eq of state}

2036: \end{equation}

2037: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2038: From equations (\ref{eq: A_Phi0_constants}) and (\ref{eq: def_Kn}), 

2039: the constant $K_n$ is related to the Lagrange multiplier $\beta$ as 

2040: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2041: \begin{equation}

2042: K_n \propto \beta^{-(n-3/2)/n},

2043: \label{Appen_B: rel Kn to beta}

2044: \end{equation}

2045: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2046: so that this constant can be used as a parameter which characterizes 

2047: the temperature of the system. However, it is not sure whether 

2048: $K_n$ itself has a role of the physical temperature, which should be 

2049: determined through the efficiency of the Carnot cycle.

2050: 

2051: 

2052: The internal energy (\ref{Appn_B: energy}) and the equation of state 

2053: (\ref{Appen_B: eq of state}) give the thermodynamic first law: 

2054: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2055: \begin{eqnarray}

2056: d'Q &=& \,dE + \, P \, dV \nonumber \\

2057:     &=& \,M^{1+1/n} \left\{\,

2058: \frac{3}{2}\, \frac{dK_n}{V^{1/n}} \, + \,

2059: \left(\frac{n-3/2}{n}\right) K_n\,\,\frac{dV}{V^{1+1/n}}

2060: \right\}\, , 

2061: \label{Appen_B: dQ}

2062: \end{eqnarray}

2063: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2064: from which adiabatic changes $d'Q = 0 $ is expressed as

2065: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2066: \begin{eqnarray}

2067: K_n\, V^{(2/3 - 1/n)} = {\rm constant}, ~~~~P\, V^{5/3} = {\rm constant'}.

2068: \label{Appen_B: Adiabatic}

2069: \end{eqnarray}

2070: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2071: Note that adiabatic lines in a $P$-$V$ plane become 

2072: steeper than isothermal ones when $n > 3/2$. 

2073: 

2074: 

2075: Now, let us consider the Carnot cycle shown in Figure \ref{fig: Carnot}. 

2076: As usual, quasi-static changes  

2077: B $\rightarrow$ C and D $\rightarrow$ A are adiabatic. As for the process 

2078: A $\rightarrow$ B, the system is in a thermal contact with a heat bath which 

2079: has a higher temperature $K_n^H$. Similarly, during the change 

2080: C $\rightarrow$ D, the system lies in a thermal equilibrium with another

2081: heat bath that has a lower temperature $K_n^L$.  

2082: The system absorbs amount of heat $Q^H$ from the higher temperature 

2083: bath and disposes $Q^L$ to the lower one during the isothermal processes 

2084: A$\rightarrow$ B and C$\rightarrow$ D, respectively. They 

2085: are easily evaluated from (\ref{Appen_B: dQ}):   

2086: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2087: \begin{equation}

2088: \begin{array}{ll}

2089: Q^H &= (n-\frac{3}{2})\,M^{1+1/n}\,\, K_n^H \,\,

2090: \Bigl(V_A^{-1/n}- V_B^{-1/n} \Bigl) ,  \\

2091: Q^L &= (n-\frac{3}{2})\,M^{1+1/n}\,\, K_n^L \,\,

2092: \Bigl(V_D^{-1/n}- V_C^{-1/n} \Bigl) . 

2093: \end{array}

2094: \label{Appen_B: heat}

2095: \end{equation} 

2096: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2097: On the other hand, a relation between the parameters of the cycle 

2098: can be obtained from the equation of state (\ref{Appen_B: eq of state}) 

2099: and the adiabatic changes (\ref{Appen_B: Adiabatic}): 

2100: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2101: \begin{eqnarray}

2102: \Biggl(\frac{K_n^H}{K_n^L}\Biggl)^{\gamma} = \frac{V_C}{V_B} = 

2103: \frac{V_D}{V_A}~;~~~\gamma = \frac{3}{2}\,\,\frac{n}{n-3/2}.

2104: \label{Appen_B: relations}

2105: \end{eqnarray}

2106: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2107: Thus, equations (\ref{Appen_B: heat}) and (\ref{Appen_B: relations}) 

2108: lead to the following efficiency of the Carnot cycle:  

2109: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2110: \begin{equation}

2111: \eta \equiv 1 - \frac{Q^L}{Q^H} =\,\, 

2112: 1 \,\,- \,\,\Biggr(\frac{K_n^L}{K_n^H}\Biggl)^{n/(n-3/2)} 

2113: =\,\, 1 \,\,- \,\,\frac{\beta^H}{\beta^L},

2114: \end{equation}

2115: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2116: where we used the relation (\ref{Appen_B: rel Kn to beta}) in the last line. 

2117: This clearly shows that the inverse of the Lagrange multiplier $\beta$ 

2118: has a role of the physical temperature. 

2119: %

2120: %

2121: %

2122: %

2123: %

2124: %

2125: \clearpage

2126: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2127: %%%%%%%%%%%%%%%%%%%%%%%%%       References     %%%%%%%%%%%%%%%%%%%%%%%%%%%%

2128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2129: %

2130: %

2131: %

2132: %

2133: %

2134: %

2135: \begin{thebibliography}{00}

2136: %

2137: %

2138: %

2139: \bibitem{T1988} C. Tsallis, J.Stat.Phys. 52 (1988) 479.

2140: \bibitem{TS2002} A. Taruya, M. Sakagami, Physica A 307 (2002) 185.

2141: \bibitem{BT1987} J. Binney, S. Tremaine, {\it Galactic Dynamics} 

2142: (Princeton Univ. Press, Princeton, 1987). 

2143: \bibitem{EPI1987} R. Elson, P. Hut and S. Inagaki, Ann. Rev. Astron. 

2144: Astrophys. 25 (1987) 565.

2145: \bibitem{MH1997} G. Meylan, D.C. Heggie, Astron.Astrophys.Rev. 8 (1997) 1.

2146: \bibitem{Antonov1962} V.A. Antonov, {\it Vest. Leningrad Gros. Univ.}, 

2147: 7 (1962) 135 (English transl. in {\it IAU Symposium 113, Dynamics of 

2148: Globular Clusters}, ed. J. Goodman and P. Hut [Dordrecht: Reidel], 

2149: pp. 525--540 [1985]) 

2150: \bibitem{LW1968} D. Lynden-Bell, R. Wood, Mon.Not.R.Astr.Soc. 138 (1968) 495.

2151: \bibitem{Padmanabhan1989} T. Padmanabhan, Astrophys.J.Suppl. 71 (1989) 651. 

2152: \bibitem{Padmanabhan1990} T. Padmanabhan, Phys.Rep. 188 (1990) 285.

2153: \bibitem{T1999} C. Tsallis, Braz. J. Phys. 29 (1999) 1.

2154: \bibitem{AO2001} S. Abe, Y. Okamoto (Eds.), {\it Nonextensive Statistical Mechanics and Its Applications} (Springer, Berlin, 2001) 

2155: \bibitem{PP1993} A.R. Plastino, A. Plastino, Phys.Lett. A 174 (1993) 384. 

2156: \bibitem{PP1999} A.R. Plastino, A. Plastino, Braz. J. Phys. 29 (1999)

2157:         79.  

2158: \bibitem{MNPP2000} S. Mart\'inez, F. Nicol\'as, F. Pennini, A. Plastino, 

2159: Physica A 286 (2000) 489.

2160: \bibitem{AMPP2001} S. Abe, S. Mart\'inez, F. Pennini, A. Plastino, 

2161: Phys.Lett. A 281 (2001) 126.

2162: \bibitem{CT1991} E.M.F. Curado, C. Tsallis, J.Phys.A 24 (1991) L69. 

2163: \bibitem{PP1997} A. Plastino, A.R. Plastino, Phys.Lett. A 226 (1997) 257.

2164: \bibitem{TMP1998} C. Tsallis, R.S. Mendes, A.R. Plastino, 

2165: Physica A 261 (1998) 534.

2166: \bibitem{Chandra1939} S. Chandrasekhar, {\it Introduction to the Study of

2167:         Stellar Structure} (New York, Dover, 1939) 

2168: \bibitem{KW1990} R. Kippenhahn, A. Weigert, {\it Stellar Structure 

2169: and Evolution} (Springer, Berlin, 1990)

2170: \bibitem{deVegaSanchez} H.J. de Vega, N. S\'anchez, 

2171: Nucl.Phys. B 625 (2002) 409.

2172: \bibitem{Chavanis2002} P.H. Chavanis, Astron. \& Astrophys. 381 (2002) 340.

2173: \bibitem{Chavanis2001} P.H. Chavanis, astro-ph/0108378.

2174: \bibitem{AR2000} S. Abe, A.K. Rajagopal, Phys.Lett. A 272 (2000) 345; 

2175:   J. Phys. A 33 (2000) 8733; Europhys. Lett. 52 (2000) 610. 

2176: %

2177: %

2178: %

2179: \end{thebibliography}

2180: %

2181: %

2182: %

2183: %

2184: %

2185: %

2186: %

2187: %

2188: %

2189: %

2190: %

2191: %

2192: %

2193: %

2194: %

2195: %

2196: %

2197: %

2198: %

2199: %

2200: %

2201: %

2202: %

2203: %

2204: %

2205: %

2206: %

2207: \clearpage

2208: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2209: %%%%%%%%%%%%%%%%%%%%       Table        %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2210: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2211: %

2212: %

2213: %

2214: %

2215: %

2216: %

2217: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2218: \begin{table}

2219: \caption{Critical values of the radius-mass-temperature relation, 

2220: $\eta_{\rm crit}$ and the density contrast between center and edge, 

2221: $D_{\rm crit}=(\rho_c/\rho_e)_{\rm crit}$ in the case of a system in contact 

2222: with a heat bath for given polytrope index $n$ or $q$.}

2223: \label{tab: ErGM}

2224: 

2225: \vspace*{0.5cm}

2226: 

2227:   \begin{center}

2228: \begin{tabular}{|ccccc|} 

2229: \hline

2230:  \makebox[1.0cm]{n}  & \makebox[1.0cm]{q} &\makebox[0.1cm]{} &

2231:  \makebox[2.5cm]{$\eta_{\rm crit}$} 

2232: & \makebox[2.5cm]{$D_{\rm crit}$}\\ 

2233: \hline\hline

2234: 3 & $\frac{5}{3}$ && ----- & ----- \\

2235: 4 & $\frac{7}{5}$ && 0.9421 & 153.5 \\

2236: 5 & $\frac{9}{7}$ && 1.193 & 88.15 \\

2237: 6 & 1.22 &&  1.379 & 68.38 \\

2238: 7 & 1.18 &&  1.520 & 58.86 \\

2239: 8 & 1.15 &&  1.631 & 53.28 \\

2240: 9 & 1.13 &&  1.720 & 49.62 \\

2241: 10 & 1.12 &&  1.793 & 47.04 \\

2242: 30 & 1.04 &&  2.263 & 35.89 \\

2243: 50 & 1.02 &&  2.363 & 34.28 \\

2244: 100 & 1.01 &&  2.440 & 33.17 \\

2245: $\infty$ & 1 &&  2.518 & 32.13

2246: \\ \hline

2247: \end{tabular}

2248:   \end{center}

2249: \end{table}

2250: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2251: %

2252: %

2253: %

2254: %

2255: %

2256: %

2257: %

2258: \clearpage

2259: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2260: %%%%%%%%%%%%%%%%%%%%       Figures        %%%%%%%%%%%%%%%%%%%%%%%%%%%%

2261: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2262: %

2263: %

2264: %

2265: %

2266: %

2267: %

2268: 

2269: \vspace*{0.5cm}

2270: 

2271: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2272: \begin{figure}[h]

2273:   \begin{center}

2274:     \includegraphics*[width=13cm]{profile.eps}

2275:   \end{center}

2276:     \caption{Density profiles of stellar polytrope for $n<5$

2277:         ({\it left}) and $n\geq5$({\it right}). }

2278:     \label{fig: profile}

2279: \end{figure}

2280: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2281: 

2282: \vspace*{0.8cm}

2283: 

2284: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2285: \begin{figure}[ht]

2286:   \begin{center}

2287:     \includegraphics*[width=10cm]{eta_lambda.eps}

2288:   \end{center}

2289:     \caption{Trajectory of Emden solutions in $(\eta,\lambda)$-plane.}

2290:     \label{fig: eta_lambda}

2291: \end{figure}

2292: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2293: 

2294: \vspace*{0.8cm}

2295: 

2296: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2297: \begin{figure}

2298:   \begin{center}

2299:     \includegraphics*[width=13.cm]{c_v.eps}

2300:   \end{center}

2301:     \caption{Normalized specific heat per particle

2302:       $C_{\rm \scriptscriptstyle V}^*/N$ 

2303:       as a function of density contrast 

2304:       $\rho_c/\rho_e$ near the critical polytrope indices 

2305:         $n=3$({\it upper}) and $n=5$({\it middle}), and 

2306:         large $n$ cases({\it lower}). Here, 

2307:         the normalized specific heat $\Cv^*$ is defined by 

2308:         $\Cv/(h^2/GMr_e)^{(3/2)/(n-3/2)}$. 

2309:         \label{fig: C_V}}

2310: \end{figure}

2311: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2312: 

2313: 

2314: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2315: \begin{figure}

2316:   \begin{center}

2317:     \includegraphics*[width=13cm]{uv_crit.eps}

2318:   \end{center}

2319:     \caption{Stability/instability criterion for a system 

2320:         in contact with a thermal bath in $(u,v)$-plane. 

2321:         The thick solid lines represent the trajectories of 

2322:         Emden solutions, while the thin-solid and dashed lines 

2323:         respectively denote the conditions (\ref{eq: criterion_1}) 

2324:         and (\ref{eq: crit_curve}). 

2325:       \label{fig: uv_crit}}

2326: \end{figure}

2327: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2328: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2329: \begin{figure}

2330:   \begin{center}

2331:     \includegraphics*[width=10cm]{carnot.eps}

2332:   \end{center}

2333:     \caption{A schematic description of Carnot cycle. 

2334:       \label{fig: Carnot}}

2335: \end{figure}

2336: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 

2337: %

2338: %

2339: %

2340: %

2341: %

2342: %

2343: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2344: \end{document}

2345: