1: \documentclass[12pt,preprint]{aastex}

2: \usepackage{emulateapj5}

3: 

4: 

5: \newcommand{\vdag}{(v)^\dagger}

6: \newcommand{\myemail}{hwlee@sejong.ac.kr}

7: 

8: \slugcomment{Not to appear in Nonlearned J., 45.}

9: 

10: \shorttitle{The Kramers-Heisenberg Formula and the Gunn-Peterson

11: Troughs}

12: 

13: \shortauthors{Bach \& Lee}

14: 

15: \begin{document}

16: 

17: \title{The Kramers-Heisenberg Formula and the Gunn-Peterson Trough

18: from the First Objects in the Universe}

19: 

20: \author{Kiehunn Bach}

21: \affil{Department of Astronomy, Yonsei University,

22:     Seoul, Korea}

23: 

24: \and

25: 

26: \author{Hee-Won Lee}

27: \affil{Department of Earth Sciences, Sejong University, Seoul,

28: Korea \\

29: {\rm hwlee@sejong.ac.kr}}

30: 

31: \begin{abstract}

32: Recent cosmological observations indicate that the

33: reionized universe may have started at around $z=6$, where a

34: significant suppression around Ly$\alpha$ has been observed from

35: the neutral intergalactic medium. The associated neutral

36: hydrogen column density is expected to exceed $10^{21}{\rm\ cm^{-2}}$,

37: where it is very important to use the accurate

38: scattering cross section known as the Kramers-Heisenberg formula that is

39: obtained from the fully quantum mechanical time-dependent second order

40: perturbation theory. We present the 

41: Kramers-Heisenberg formula and compare

42: it with the formula introduced in a heuristic way by Peebles (1993)

43: treating the hydrogen atom as a two-level atom, from which we find a devitaion

44: by a factor of two in the red wing region far from the line center.

45: Adopting simple cosmological models, we compute the Gunn-Peterson 

46: optical depths and the trough profiles.

47: Our results are compared with the works performed by Madau \& Rees

48: (2000), who adopted the cross section

49: introduced by Peebles (1993). We find deviations up to 5 per cent in the

50: Gunn-Peterson transmission coefficient for an accelerated expanding universe 

51: in the red off-resonance 

52: wing part with the rest wavelength $\Delta\lambda\sim 10{\rm\ \AA}$.

53: \end{abstract}

54: 

55: 

56: \keywords{cosmology: theory --- intergalactic medium

57: -- quasars: absorption lines --- radiative transfer --- 

58: galaxies: high-redshift }

59: 

60: \section{Introduction}

61: 

62: The quasar absorption systems have been excellent tools to

63: investigate the intergalactic medium (IGM), from which

64: it has been well known that the IGM of the nearby universe is highly

65: ionized (e.g. Peebles 1993). Since the universe after the

66: recombination era $z\sim 1100$ should be dominantly neutral, there

67: must be some epoch when the universe began to be re-ionized.

68: Intensive studies have been performed on the emergence of the

69: first objects that ended the dark age of the universe. Numerical

70: calculations adopting the cold dark matter models predicted

71: the reionization epoch in $z\sim 6-12$ (e.g. Gnedin \& Ostriker

72: 1997).

73: 

74: Around this epoch a broad absorption trough in the blue part of

75: Ly$\alpha$ is expected and regarded as a strong indicator of

76: the reionization of the universe, which was predicted by Gunn \& Peterson 

77: (1965) and independently  also by Scheuer (1965).

78:  With the advent of the Hubble Space Telescope and 8 meter class

79: telescopes there have been extensive searches for the

80: Gunn-Peterson trough in the spectra of high red shift objects.

81: Remarkable contributions are made by the Sloan Digital Sky Survey, 

82: from which a number of high red shift quasars with $z$

83: ranging from 4 to 6 have been found. According to the recent

84: report from the Keck spectroscopy of these high red shift quasars

85: (Becker et al. 2001), the flux level drop around Ly$\alpha$ is

86: much higher for the quasar with $z=6.28$ than those for other

87: quasars with $z<6$, which indicates that the reionization epoch

88: may be found at around $z\sim 6$.

89: 

90: The exact computation of the flux drop around Ly$\alpha$  requires

91: an accurate atomic physical estimation of the scattering cross

92: section. Recent theoretical works on the calculation of the

93: Gunn-Peterson trough were provided by 

94: Miralda-Escud\'e (1998) and Madau \& Rees (2000), who adopted

95: the formula that was introduced in a heuristic way using the second

96: order time-dependent perturbation theory by Peebles (1993). The

97: formula is derived based on the assumption that the hydrogen atom

98: is a two-level atom, in order to show the behavior of the

99: scattering cross section that is approximated by the Lorentzian

100: near resonance and yields $\omega^4$ dependence in the low energy

101: limit. As Peebles noted clearly in the text, due to the two-level

102: assumption, the formula provides an inaccurate proportionality

103: constant in the low energy limit, even though it correctly gives

104: the $\omega^4$ dependence.

105: 

106: The accurate cross section should be obtained from the second order

107: time-dependent perturbation theory treating the hydrogen atom as

108: an infinitely many-level atom including the continuum free states, which

109: is known as the Kramers-Heisenberg formula. 

110: The discrepancy between the two formulae will be significant in the far

111: off-resonance regions where the contribution from the $np (n>2)$ states 

112: including the continuum states becomes considerable. Therefore, 

113: in order to obtain an accurate Gunn-Peterson profile

114: it is essential to investigate the exact scattering 

115: optical depth of a medium with a high neutral hydrogen column density.

116: 

117: In this Letter, we present a faithful atomic physics that governs

118: the scattering around Ly$\alpha$ by introducing the Kramers-Heisenberg 

119: formula and a simple fitting formula around Ly$\alpha$.  We compute the

120: Gunn-Peterson trough profiles adopting a representative set of

121: cosmological parameters with our choice of the reionization epoch

122: and make quantitative comparisons with previous works.

123: 

124: \section{The Kramers-Heisenberg Formula}

125: 

126: The interaction of photons and electrons is described by the

127: second order time-dependent perturbation theory, of which the

128: result is summarized as the famous Kramers-Heisenberg formula. As

129: is well illustrated in a typical quantum mechanics text, it may

130: be written as

131: 

132: \begin{eqnarray}

133: {d\sigma\over d\Omega}(\omega) &=& {r_0^2\over m_e^2\hbar^2}\Bigg\vert

134: \sum_{I}\left[ {\omega(\vec p\cdot

135: \hat\epsilon^{(\alpha')})_{AI}(\vec p\cdot

136: \hat\epsilon^{(\alpha)})_{IA}

137: \over

138: \omega_{IA}(\omega_{IA} -

139: \omega - i\Gamma_I/2)} \right. \nonumber \\

140: &-&\left. {\omega(\vec p\cdot

141: \hat\epsilon^{(\alpha)})_{AI}(\vec p\cdot

142: \hat\epsilon^{(\alpha')})_{IA}

143: \over

144: \omega_{IA}(\omega_{IA} +

145: \omega)}\right]\Bigg\vert^2

146: \end{eqnarray}

147: where $\hat\epsilon^{\alpha}, \hat\epsilon^{\alpha'}$ are the polarization

148: vectors associated with the incident photon and outgoing photon, respectively,

149: $\Gamma_I$ is the radiation damping term associated with the intermediate 

150: state $I$, $\omega$ is the incident angular frequency, $\omega_{IA}$ is

151: the angular frequency of the transition between I and the ground state $A$,

152: and $r_o={e^2/(m_e c^2)}= 2.82 \times 10^{-13}{\rm\ cm}$ is the

153: classical electron radius (e.g. Sakurai 1967). 

154: 

155: Here, the electron is in the ground

156: state $A$ before scattering and de-excites to the same state. The

157: summation (and integration) should be carried over all the

158: intermediate states $I$ including the infinite number of bound

159: states and the free or continuum states. 

160: For hydrogen, the dipole moment matrix elements have been explicitly

161: given using the recurrence relations of the hypergeometric

162: function in many texts and in the literature (e.g.

163: Berestetski, Lifshitz \& Pitaevskii 1971, Bethe \& Salpeter 1957, 

164: Karzas \& Latter 1961).

165: 

166: In the blue part of Ly$\alpha$, it is possible that the scattering

167: atom may de-excite to the excited $2s$ state by re-emitting a

168: photon with much lower frequency than the incident photon. This

169: inelastic scattering or the Raman scattering is negligible near

170: Ly$\alpha$ due to small phase space available for an outgoing

171: photon. However, this process becomes important as the incident

172: photon energy increases. In the range where the current work is

173: concerned, the Raman scattering process is safely neglected.

174: 

175: Despite the existence of the explicit analytic expressions of each

176: matrix element that constitutes the Kramers-Heisenberg formula for

177: hydrogen, it is still cumbersome to use the formula as it is.

178: Therefore, a simple fitting formula around Ly$\alpha$ will be 

179: useful for practical applications. Near resonance ($1170{\rm\

180: \AA}\ <\lambda<1410{\rm\ \AA}$), the Lorentzian function gives

181: quite a good approximation

182: \begin{equation}

183: \sigma(\omega)=\frac{3\lambda_\alpha^2}{8 \pi}

184: \frac{\Gamma_{2p}^2} {(\omega-\omega_\alpha)^2 + \Gamma_{2p}^2 /4},

185: \end{equation} 

186: where $\Gamma_{2p}=6.25\times10^8{\rm\ s^{-1}}$ is the radiation

187: damping constant associated with the Ly$\alpha$ transition.

188: In the long wavelength region

189: ($\lambda>1410{\rm\ \AA}$), Gavrila (1967) provided the fitting

190: polynomial for the Rayleigh scattering cross section, which is

191: \begin{eqnarray}

192: \sigma(\omega)/\sigma_T &=&

193: 0.400 (\omega/\omega_\alpha)^4 

194: + 0.900 (\omega/\omega_\alpha)^6  \nonumber \\

195: &+& 12.6 (\omega/\omega_\alpha)^{14}.

196: \end{eqnarray}

197: In particular, Ferland (2001)

198: applied Gavrila's fit to his photoionization code `Cloudy'.

199: 

200: In the case of the short wavelength region 

201: ($1070{\rm\ \AA}\ <\lambda<1170{\rm\ \AA}$), 

202: we provide a similar polynomial fit to

203: the Kramers-Heisenberg formula

204: \begin{eqnarray}

205: \sigma(\omega)/\sigma_T &=&1.62\times10^6 (\omega_\alpha/\omega)^4

206: +5.88\times10^6 (\omega_\alpha/\omega)^3 \nonumber \\

207: &+&7.99\times10^6 (\omega_\alpha/\omega)^2

208: -4.83\times10^6 (\omega_\alpha/\omega) \nonumber \\

209: &+&1.09\times10^6.

210: \end{eqnarray}

211: The deviation of the fit is within 5 per cent from the true

212: Kramers-Heisenberg formula. 

213: 

214: In Fig.~1 we show the scattering cross section from the

215: Kramers-Heisenberg formula by the solid line and by the dotted

216: line we represent the fit. The behavior near resonance is depicted in the

217: bottom panel, because the cross section changes very steeply. 

218: It is apparent that the scattering

219: cross section is excellently approximated by the Lorentzian. 

220: However, in the wavelength range considered in this Letter, 

221: the radiation damping is completely 

222: negligible, and the curve shown in the figure is simply proportional

223: to $\Delta\omega^{-2}=(\omega-\omega_{Ly\alpha})^{-2}$.  The deviation 

224: is slightly anti-symmetric with respect to the line center in the

225: sense that the cross section in the blue part is smaller than the Lorentzian

226: and in the red part it is larger than the Lorentzian. 

227: 

228: Therefore, 1 per

229: cent of deviation of the Lorentzian from the Kramers-Heisenberg formula

230: is seen at a wavelength shift

231: of $\Delta\lambda=\lambda-\lambda_\alpha = \pm 3.3{\rm\ \AA}$, for which

232: the corresponding cross section $\sigma =3.8\times 10^{-21}{\rm\ cm^2}$.

233: This indicates the accuracy of the Voigt profile fitting applied to

234: quasar absorption systems, where the

235: accuracy is more than 99 per cent when the absorbing medium is

236: characterized with the H~I column density smaller than $3\times 10^{20}

237: {\rm\ cm^{-2}}$.

238: 

239: Further away from

240: the line center, the  cross section in the blue part decreases very 

241: steeply till $\lambda\sim 1100 {\rm\ \AA}$, but the decrease of the cross

242: section in the red part is rather gradual and eventually becomes 

243: proportional to $\omega^4$, which corresponds to the classical result.

244: 

245: Fig.~1 also shows the comparison of the Kramsers-Heisenberg formula 

246: and the heuristic formula 

247: \begin{equation}

248: \sigma_P(\omega) = {3\lambda_\alpha^2\over 8\pi}

249: {\Gamma_{2p}^2 (\omega/\omega_\alpha)^4

250: \over

251: {(\omega-\omega_\alpha)^2+\Gamma_{2p}^2(\omega/\omega_\alpha)^6/4}}

252: \end{equation}

253: introduced by Peebles (1993).  

254: 

255: The $\omega^4$ dependence in the limit $\omega\ll\omega_\alpha$ 

256: is obtained as a result of the closure relation, which is apparent in the both 

257: formulae.  However,

258: the Kramers-Heisenberg formula gives about twice larger scattering

259: cross section than $\sigma_P$ does. This deviation is easily

260: noted when the oscillator strength of the Ly$\alpha$ transition is

261: $f_{Ly\alpha}=0.42$. The scattering cross section in the far red

262: region is contributed from all the $p$ states and the oscillator

263: strength is a good measure of the contributions of each individual

264: excited state. This implies that the $2p$ state contribution is

265: comparable to the total contributions from the remaining states in

266: the low energy limit.

267: 

268: \section{Gunn-Peterson Trough Profiles with the

269: Kramers-Heisenberg formula}

270: 

271: We compute the Gunn-Peterson optical depth defined by

272: \begin{equation}

273: \tau_{GP}=\int_{z_{rei}}^{z_s}\ dz\ {dl\over dz}\sigma[\nu=

274: c(1+z)/\lambda_{obs}]\ n(z),

275: \end{equation}

276: with the Kramers-Heisenberg formula and make comparisons with

277: previous works performed by Madau \& Rees (2000) (see also

278: Miralda-Escud\'e 1998). Here, $\lambda_{obs}$ is the observed wavelength,

279: $z_{rei},\ z_s$ are the redshifts of the complete reionization of the

280: universe and the reionizing source, and $n(z)=n_0(1+z)^3$ is the homogeneous

281: neutral hydrogen density at redshift $z$.  We choose

282: $z_{rei}=6,\ z_{s}=7$ as in Madau \& Rees (2000), but do not consider 

283: the proximity effect of the ionizing source. The Gunn-Peterson optical 

284: depth can be written as

285: \begin{eqnarray}

286: \tau_{GP} &=& N_{0\, HI} \int_{z_{rei}}^{z_s}\ dz\ 

287: \sigma[\nu=c(1+z)/\lambda_{obs}] \nonumber \\

288: &\times& (1+z)^2/[\Omega_M(1+z)^3 +\Omega_\Lambda]^{1/2},

289: \end{eqnarray}

290: where $\Omega_M, \Omega_\Lambda$ are the density parameters due to matter

291: and the cosmological constant and the characteristic hydrogen column density 

292: \begin{equation}

293: N_{0\, HI}\equiv n_0\ c\ H_0^{-1}.

294: \end{equation}

295: We choose the present Hubble constant and the hydrogen number density

296: $H_0=50{\rm\ km\ s^{-1}\ Mpc^{-1}},\ 

297: n_0=2.4\times10^{-7}{\rm\ cm^{-3}}$ so that $N_{0\, HI}=4.3\times

298: 10^{21}{\rm\ cm^{-2}}$.

299: 

300: In Fig.~2, we show the Gunn-Peterson transmission coefficient $T_{GP}\equiv

301: e^{-\tau_{GP}}$.  We plot $T_{GP}$ for the case $\Omega_M=1, 

302: \Omega_\Lambda=0$ in the top panel and the same quantity 

303: for an accelerated expanding

304: universe $\Omega_M=0.35, \Omega_\Lambda=0.65$ in the bottom panel. 

305: In terms of the characteristic Gunn-Peterson

306: optical depth $\tau_{0\, GP}$ defined as

307: \begin{equation}

308: \tau_{0\, GP}(z_s)\equiv {3\lambda_\alpha^3\Gamma_{2p} n(z_s)

309: \over 8\pi H(z_s)},

310: \end{equation}

311: our choice of parameters in the case of the top panel yields 

312: the value of $\tau_{0\, GP}=3\times 10^5$ at $z_s=7$ as was adopted in the work 

313: of Madau \& Rees (2000).

314: 

315: The horizontal axis represents the logarithm of the normalized wavelength 

316: ratio $\delta$ defined as

317: \begin{equation}

318: \delta \equiv {\lambda_{obs}\over \lambda_\alpha(1+z_s)}-1,

319: \end{equation}

320: from which $\delta=0$ corresponds to the resonance wavelength of Ly$\alpha$.

321: By the dotted lines

322: we present the Gunn-Peterson transmission coefficient obtained using the

323: scattering cross section given by Eq.~(5). The deviation between the

324: two formulae is notable around $\delta =10^{-2}$, where 

325: the deviation is about 3 per cent in the top panel and 5 per cent 

326: in the bottom panel. The Peebles approximation turns out to be pretty 

327: good for contemporary application.

328: 

329: The deviation between the two formulae will increase as $n_0$ or

330: $N_{0\, HI}$ increases, because the discrepancy of the Kramers-Heisenberg 

331: formula and Eq.~(5) becomes larger as the frequency is further away from

332: the line center. Near resonance, both formulae are excellently approximated

333: by the same Lorentzian. Therefore, no significant deviation is expected

334: when the neutral medium is of low column density $\la 10^{21}{\rm\ cm^{-2}}$.

335: It is notable that an accurate treatment of atomic physics is more important

336: in an accelerated expanding universe where the

337: univere was more compact than the universe without the cosmological constant.

338: 

339: 

340: \section{Summary and Discussion}

341: 

342: In this Letter, we have investigated the behavior of the scattering

343: cross section around Ly$\alpha$ in a quantitative way, where the deviation 

344: from the Lorentzian becomes significant as the incident frequency gets further

345: away from the line center. Therefore, in an analysis of the Gunn-Peterson

346: trough profile, which is associated with a neutral medium with a high H~I

347: column density, an inaccurate treatment of

348: the atomic physics of hydrogen may introduce significant errors in estimating

349: important cosmological parameters including the epochs of the emergence

350: of the first objects and the completion of the reionization of the universe.

351: 

352: Voigt profile fitting has been very successfully applied to

353: quasar absorption systems with a broad range of H~I column densities. 

354: However, the deviation of the true scattering cross

355: section from the Lorentzian exceeds 1 per cent

356: when the relevant column density becomes $N_{HI}\ge 3\times

357: 10^{20} {\rm\ cm^{-2}}$ that is the typical column density of a

358: damped Ly$\alpha$ absorber.  This is especially

359: important in some damped Ly$\alpha$ systems that may possess $N_{HI} >

360: 10^{21}{\rm\ cm^{-2}}$ (e.g. Turnshek \& Rao 1998). However, it should

361: be noted that the damping constant $\Gamma_{2p}$ is so small compared

362: with the scale relevant in this work, the cross section is effectively of the

363: form $\propto\Delta\omega^{-2}=(\omega-\omega_\alpha)^{-2}$. Therefore, the

364: absorption profile is irrelevant to the exact value of the 

365: radiation damping term, which means that the term `damped Ly$\alpha$ 

366: absorption' is a misnomer.

367: 

368: 

369: \acknowledgments

370: KHB thanks the support from the BK21 project

371: initiated by the Ministry of Education. HWL thanks Roger Blandford and

372: Sang-Hyeon Ahn for their helpful discussions.

373: 

374: \begin{thebibliography}{}

375: \bibitem[Becker et al. 2001]{bec01} Becker, R. H., Fan, X., White, R. L., 

376: Strauss, M.  A., Narayanan, V. K., 2001, \aj \ submitted (astro-ph/0108097)

377: \bibitem[Berestetskii et al. 1971]{ber71} Berestetskii, V. B., Lifshitz, E. M.,

378:  \& Pitaevskii, L. P., 1971, Relativistic Quantum Theory, Pergamon Press, 

379: Oxford

380: \bibitem[Bethe \& Salpeter 1957]{bet57} Bethe, H. A., Salpeter, E. E., 1957, 

381: Quantum Mechanics of One and Two Electron Atoms, Academic Press Inc. New York

382: \bibitem[Ferland 2001]{fer01} Ferland, G., {\it Hazy, a brief introduction

383: to Cloudy 94.00}

384: \bibitem[Gavrila 1967]{gav67} Gavrila, M., 1967, Phys. Rev., 163, 147 

385: \bibitem[Gnedin \& Ostriker 1997]{gne97} Gnedin, N. Y., \& Ostriker, J. P., 

386: 1997, \apj, 486, 581

387: \bibitem[Gunn \& Peterson 1965]{gun65} Gunn, J. E., Peterson, B. A., 1965, 

388: \apj, 142, 1633

389: \bibitem[Karzas \& Latter 1961]{kar61} Karzas, W. J., \& Latter, R., 1961,

390: \apjs, 6, 167

391: \bibitem[Madau \& Rees 2000]{mad00} Madau, P., Rees, M. J., 2000, \apj, 542, 

392: L69

393: \bibitem[Miralda-Escud\'e 1998]{mir98} Miralda-Escud\'e, J., 1998, \apj, 

394: 501, 15

395: \bibitem[Peebles 1993]{peb93} Peebles, P. J. E., 1993, 

396: Principles of Physical Cosmology, Princeton University Press, Princeton

397: \bibitem[Sakurai 1967]{sak67} Sakurai, J. J., 1967, Advanced Quantum Mechanics,

398: Addison-Wesley Publishing Company, Reading, Massachusetts

399: \bibitem[Scheuer 1965]{sch65} Scheuer, P. A. G. 1965, Nature, 207, 963

400: \bibitem[Turnshek \& Rao 1998]{tur98} Turnshek, D. A., \& Rao, S., 1998, ApJ, 

401: 500, L115

402: \end{thebibliography}

403: 

404: 

405: \clearpage

406: 

407: 

408: \begin{figure}

409: \plotone{f1.eps} \caption{

410: The scattering cross section around Ly$\alpha$. The solid line represents

411: the accurate, fully quantum mechanical cross section known as the 

412: Kramers-Heisenberg formula. The dotted line represents the Lorentzian

413: given by Eq.~(2), which gives an excellent approximation near the line

414: center. The long dashed line represents the fitting formula provided by

415: Gavrila (1967), of which the approximation is valid for $\lambda>1400{\rm

416: \ \AA}$. The dot-dash line represents our fit to the Kramers-Heisenberg

417: formula in the blue part given in Eq.~(4). The dot-long dash line represents

418: the cross section obtained from Eq.~(5),

419: which is inaccurate by a factor of two in the far red wing region

420: and accurate near the line center. In the bottom panel, we plot the cross

421: section obtained from the Kramers-Heisenberg formula (solid line) and 

422: the Lorentzian Eq.~(2).  The cross section is asymmetric relative 

423: to the line center in the sense that the cross section in the blue part 

424: is smaller than in the red part.

425: \label{fig1}}

426: \end{figure}

427: 

428: \clearpage

429: 

430: \begin{figure}

431: \plotone{f2.eps} \caption{

432: The Gunn-Peterson transmission coefficient $T_{GP}\equiv e^{-\tau_{GP}}$

433: for $\Omega_M=1, \Omega_\Lambda=0$ (top panel) and for

434: $\Omega_M=0.35, \Omega_\Lambda=0.65$ (bottom panel). 

435: The solid lines represent the values obtained using the Kramers-Heisenberg

436: formual and the dotted lines are for the values from the cross section

437: $\sigma_P$ introduced by Peebles (1993).

438: The present Hubble constant and the hydrogen number density are chosen to be

439: $H_0=50{\rm\ km\ s^{-1}\ Mpc^{-1}},

440: n_0=2.4\times10^{-7}{\rm\ cm^{-3}}$ so that $N_{0\, HI}=n_0 c H_0^{-1}=

441: 4.3\times 10^{21}{\rm\ cm^{-2}}$.

442: \label{fig2}}

443: \end{figure}

444: 

445: 

446: \end{document}

447: