1: %\documentstyle[twocolumn,aps]{revtex}

2: %\documentstyle[preprint,aps]{revtex}

3: \documentstyle[prl,aps,amsfonts,amssymb,twocolumn,epsfig]{revtex}

4: \begin{document}

5: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname @twocolumnfalse\endcsname

6: \draft

7: \title{Plasma-Induced Frequency Chirp of Intense Femtosecond Lasers \\

8:        and Its Role in Shaping High-Order Harmonic Spectral Lines}

9: \author{Jung-Hoon Kim and Chang Hee Nam}

10: \address{Department of Physics and Coherent X-Ray Research Center,

11:          Korea Advanced Institute of Science and Technology,

12:          Taejon 305-701, Korea}

13: \date{\today}

14: \maketitle

15: 

16: \begin{abstract}

17: We investigate the self-phase modulation of intense femtosecond

18: laser pulses propagating in an ionizing gas and its effects on

19: collective properties of high-order harmonics generated in the 

20: medium. Plasmas produced in the medium are shown to induce a positive

21: frequency chirp on the leading edge of the propagating laser pulse,

22: which subsequently drives high harmonics to become positively 

23: chirped. In certain parameter regimes, the plasma-induced positive 

24: chirp can help to generate sharply peaked high harmonics, by 

25: compensating for the dynamically-induced negative chirp that is 

26: caused by the steep intensity profile of intense short laser pulses.

27: \end{abstract}

28: 

29: \pacs{42.65.Ky, 52.40.Nk, 32.80.Rm}

30: 

31: \vskip1.5pc]

32:  

33: 

34: Recent remarkable progress in high-power femtosecond laser 

35: technology has provided a novel opportunity to investigate 

36: high-order harmonic generation processes in an unprecedented 

37: high-intensity, ultrashort pulse regime. While necessary for 

38: generating high-order harmonics with wavelength extremely short 

39: compared to the laser wavelength\cite{Spiel,Chang}, the use of 

40: lasers in this regime has some other important merits: the harmonic

41: conversion efficiency is dramatically enhanced\cite{Kan1,Schn}, and

42: the harmonic pulse duration can be reduced to a subfemtosecond time

43: scale\cite{Schaf,Lee}. 

44: 

45: As the laser pulse evolves into the high-intensity, ultrashort 

46: regime, high harmonic emission begins to exhibit complicated

47: spectral features, such as harmonic line broadening, blueshifting, 

48: splitting, and smearing\cite{Schaf,Lee,Watson,Kan2,Shin,Kim1}. 

49: All of these features are absent in the high harmonic spectra 

50: observed for lasers with weak intensity well below the ionization

51: saturation intensity and long pulse duration (longer than a few 

52: hundred femtoseconds) that show well-defined narrow harmonic peaks

53: almost exactly at odd multiples of the fundamental laser frequency.

54: The involved spectral features showing up for intense short laser 

55: pulses can be satisfactorily accounted for, at least at the 

56: single-atom level, in terms of decomposition of high harmonics 

57: into quantum-path components, and different behaviors of the 

58: quantum-path components, in laser fields with rapidly-varying 

59: instantaneous intensity\cite{Kim1,Gaarde,Balcou}.

60: 

61: Since high-intensity laser pulses can ionize atoms producing 

62: plasmas that can affect propagating waves, a proper understanding

63: of macroscopic high harmonic structure necessitates a detailed 

64: knowledge of the plasma-related propagation effect as well as the

65: single-atom effect. It is known that an increase in electron 

66: density during the laser pulse duration causes the refractive 

67: index of a medium to decrease with time, which leads to frequency

68: upshifting and spectral broadening of the laser 

69: pulse\cite{Yablo,Wilks,Wood,Rae,Chessa}. Because the change 

70: in the driving laser spectrum should be reflected in the harmonic 

71: conversion, one might expect that harmonic spectral lines would 

72: likewise be blueshifted and broadened. In reality, however, the 

73: harmonic spectrum exhibits behavior more complicated than this 

74: expectation, as demonstrated in this paper. As a matter of fact, 

75: there is another source that can affect the harmonic spectral line

76: shape: the dynamically-induced harmonic chirp (dynamic chirp) that

77: is brought about by a steep pulse envelope\cite{Gaarde,Kim2}. 

78: The observed spectral behavior of high harmonics can only be 

79: understood through inspecting the temporal variation of harmonic 

80: frequency caused by the plasma-induced change in the laser 

81: frequency, in close connection with the dynamic chirp.

82: 

83: In this paper, we elucidate how the plasma effect modulates the 

84: spectrum of an intense femtosecond laser pulse, and discuss its 

85: subsequent influence and the effect of laser focusing on 

86: macroscopic high harmonic spectra. It is shown that the plasmas 

87: induce a positive frequency chirp on the leading edge of the laser

88: pulse up to the point at which a maximum blueshift is attained, 

89: and a negative chirp on the remaining part of the laser pulse. 

90: Depending on the relative amount of the plasma-induced chirp 

91: compared to the dynamic chirp, not only broadening but also 

92: narrowing can occur in high harmonic spectral lines. We 

93: demonstrate these using a one-dimensional (1D) model, in which 

94: the atomic response to the laser is calculated from the 1D 

95: Schr\"{o}dinger equation, and propagations of the laser and 

96: harmonic fields are considered in 1D space along the propagation

97: axis.

98: 

99: In order to see how the laser field $E_1$ is affected by the plasmas

100: (whose effects are dominated by electrons) produced in the medium, 

101: we begin by finding a solution of the 1D wave equation:

102: \begin{equation}

103: \frac{\partial ^2E_1(x,t)}{\partial x ^2} -\frac{1}{c^2}

104: \frac{\partial ^2E_1}{\partial t ^2} =

105: \frac{\omega ^2_p(x,t)}{c^2}E_1,

106: \end{equation}

107: where $\omega _p(x,t)= \omega _0[N_e(x,t)/N_{cr}]^{1/2}$ is the 

108: local plasma frequency, and $\omega _0$ is the laser frequency. The

109: critical plasma density $N_{cr}$ is given in Gaussian units by 

110: $N_{cr}=m_e\omega _0^2/4\pi e^2$, where $m_e$ is the electron mass.  

111: To calculate the electron density $N_e(x,t)$, we use the ADK 

112: model\cite{Ammo}, and consider sequential tunneling ionization up 

113: to as high

114: stages of ionization as needed, neglecting collisional ionization 

115: that is of little significance in the parameter regions in which

116: high harmonic generation experiments are commonly carried out. 

117: At the present gas pressures ($\leq$ 100 Torr) much lower than 1 

118: atm, the energy loss and temporal broadening of the laser are 

119: negligible\cite{Wood}; thus, we may ignore the amplitude modulation

120: of the laser field. Assuming the Gaussian incident pulse 

121: $E_1(x,t)=E_0\exp{[-(2\ln{2}/\Delta t^2) (t-x/c)^2 -iw_0(t-x/c)]}$

122: that is a solution to Eq.\ (1) in free space with $\Delta t$ being

123: the full width at half maximum (FWHM) of the pulse, we may then 

124: write the solution of Eq.\ (1) in the medium as

125: \begin{eqnarray}

126: E_1(x,t)&=&E_0\exp{[-(2\ln{2}/\Delta t^2)(t-x/c)^2 -i\omega_ 0t }

127: \nonumber \\

128:         & & +i\int^x n(x',t-|x-x'|/c)\frac{\omega _0}{c}dx'],

129: \end{eqnarray}

130: where $n(x,t)=[1-N_e(x,t)/N_{cr}]^{1/2}$ is the refractive index

131: of the medium. This expression turns out

132: to be a good approximate solution of Eq.\ (1) under the

133: conditions $N_e/N_{cr} \ll 1$ and $|c^{-1}\int^x(\partial n/

134: \partial t)dx'| \ll 1$, which are satisfied in the parameter 

135: regions considered in this paper.

136: 

137: To confirm that Eq.\ (2) indeed closely approximates the exact 

138: solution, we present in Fig.\ 1 some typical spectra of 30-fs

139: laser pulses at the exit of the medium (Ne gas) calculated 

140: from Eq.\ (2), along with those obtained from direct numerical 

141: calculations of Eq.\ (1). Use of the explicit expression in 

142: Eq.\ (2) of course significantly reduces the computational time. 

143: Upon comparison, it is obvious that the approximate solutions 

144: presented in Fig.\ 1(a) agree well with the exact numerical 

145: results in Fig.\ 1(b) in the parameter regions considered. It 

146: can be seen that, as the gas density increases and/or as the 

147: laser intensity increases, the spectrum shifts toward a 

148: higher-frequency region and becomes broader. 

149: 

150: More detailed features of the laser pulse passing through the 

151: medium can be revealed with the help of the Wigner distribution 

152: function\cite{Kim2}, that allows a view of temporal variation of

153: the laser spectrum. The Wigner distributions, calculated under 

154: the same conditions as in Fig.\ 1, are displayed in Fig.\ 2. It 

155: can be observed that, owing to the phase modulation induced by 

156: plasmas produced in the medium, the laser frequency increases 

157: with time (becomes positively chirped) in the leading edge, and 

158: then decreases (becomes negatively chirped) back to the original

159: frequency in the remaining part of the pulse. We note that at 

160: the moment when the plasma-induced chirp changes sign from 

161: positive to negative, the production rate of electrons reaches 

162: its maximum. This moment, at which the laser experiences a 

163: maximum blueshift, comes earlier in time as the laser intensity

164: increases, as can be seen from Fig.\ 2.  

165: 

166: When focused laser beams are used, as in usual high harmonic 

167: generation experiments, due regard should also be paid to the 

168: focusing effect that can change the amplitude and phase of the 

169: laser field. In fact, in three-dimensional (3D) simulations this 

170: focusing effect could have been automatically considered. 

171: Apparently, however, the 1D wave equation in Eq.\ (1) cannot 

172: deal with this effect. Nevertheless, via Eq.\ (2) the focusing 

173: effect can be taken into account along the propagation axis as 

174: follows:

175: \begin{equation}

176: E_{1f}(x,t)=E_1(x,t)f(x)\exp{[i\Phi _f(x)]},

177: \end{equation}

178: where $f(x)=[1+4(x/b)^2]^{-1/2}$ and $\Phi _f(x)=

179: -\tan ^{-1}(2x/b)$ represent, respectively, the amplitude and 

180: phase changes due to the focusing\cite{Sieg}, and $b$ is the 

181: confocal parameter. The laser field given in the form of Eq.\ (3)

182: now suitably describes the phase modulation induced by plasmas 

183: and the focusing effect along the axis, and can be used for

184: discussing the phase matching issue of high harmonics generated 

185: {\em on the propagation axis}.    

186: 

187: We next discuss the propagation of high harmonic fields. Since the 

188: change in the refractive index in the presence of plasmas is much

189: smaller for higher-frequency waves, we may ignore here the plasma 

190: effect. Then the Green function method enables us to write a 

191: solution of the 3D wave equation $\nabla ^2E_h -c^{-2}(\partial 

192: ^2E_h/\partial t ^2) = 4\pi c^{-2}(\partial ^2P/\partial t ^2)$

193: in the following integral form:

194: \begin{equation}

195: E_h({\bf r},t)=-\int \frac{1}

196: {c^2|{\bf r}-{\bf r}'|}

197: \frac{\partial ^2 P({\bf r}',t')}{\partial t'^2}

198: d^3r',

199: \end{equation}

200: where $t'$ is the retarded time defined by 

201: $t'=t-|{\bf r}-{\bf r}'|/c$, and the laser-induced polarization 

202: $P$ is given in terms of the gas density $N_0$ and the atomic 

203: dipole moment $d$ by $P({\bf r}',t')=N_0d({\bf r}',t')$.

204: Equation (4) clearly indicates that the harmonic field $E_h$ at 

205: $t$ is determined by the coherent sum of the dipole accelerations

206: of atoms in the medium calculated at the retarded time $t'$.

207: 

208: In our 1D calculations, a medium lies on the propagation axis, 

209: and the integration in Eq.\ (4) is performed in practice over

210: the 1D space along the axis. The medium, a Ne gas of length 

211: $l= 700$ $\mu$m at $\simeq$ 28 Torr, is uniformly discretized by

212: 200 points, and at each point the dipole acceleration is 

213: calculated by numerically solving the 1D Schr\"{o}dinger equation

214: for an atom in the laser field $E_{1f}$ given in Eq.\ (3), with 

215: $\Delta t=30$ fs, $\lambda = 800$ nm, $b=4$ mm, and $I=1\times 

216: 10^{15}$ W/cm$^2$ at the entrance of the medium. After suitably 

217: weighted by a constant factor to yield the correct gas density, 

218: the results are then added according to Eq.\ (4) to give $E_h$. 

219: 

220: The high harmonic spectrum calculated in the above way is 

221: presented in Fig.\ 3, which shows the respective roles played

222: by the laser focusing and plasmas in the formation of macroscopic

223: harmonic fields. In Fig.\ 3(b), we neglect the plasma effect by 

224: setting $n(x,t)=1$ to concentrate only on the focusing effect. 

225: Both the focusing and plasma effects are fully considered in 

226: Fig.\ 3(c), and for comparison we present in Fig.\ 3(a) a 

227: single-atom spectrum calculated for an atom located at the 

228: entrance of the medium. Whereas the single-atom spectrum in 

229: Fig.\ 3(a) is smeared with a complicated structure in the plateau

230: region, the macroscopic harmonic spectrum in Fig.\ 3(b) exhibits 

231: discrete harmonic peaks. This is because the intensity-sensitive 

232: harmonic components (the long quantum-path component and 

233: multiple-recollision components), which give rise to the 

234: complicated structure, are suppressed due to their poor 

235: phase-matching conditions, leaving only the short quantum-path 

236: component\cite{Kim1,Sali}. Here we emphasize that the variation 

237: in the laser intensity caused by the focusing as represented by 

238: $f^2(x)$ is mainly responsible for this cleaning up of the 

239: harmonic spectrum. At very low gas densities, we may observe a 

240: harmonic spectrum similar to that shown in Fig.\ 3(b). However, 

241: at the gas density used in Fig.\ 3(c), a further change in the 

242: harmonic spectrum is caused by the plasma effect: harmonics of 

243: low order (43rd and below) undergo  spectral broadening, while 

244: higher-order harmonics get sharpened. 

245: 

246: The plasma-induced harmonic line broadening and narrowing are 

247: detailed in Figs.\ 4 and 5, respectively.  The mechanism that 

248: shapes high harmonic lines can be explained by making a 

249: comparison between the plasma-induced chirp and the 

250: dynamically-induced chirp\cite{Schaf,Gaarde,Kim2}. Considering 

251: that only the leading edge of the laser pulse, where the 

252: depletion of neutral atoms is not severe and the electron density

253: is low, is important for the phase-matched harmonic generation, 

254: we focus only on this part of the laser pulse. The leading edge 

255: of the laser pulse, in which the laser frequency increases with 

256: time due to the plasma effect (Fig.\ 2), tends to generate 

257: positively chirped high harmonics. This effect is opposed by the 

258: dynamic single-atom effect that, in the present case, forces high

259: harmonics to become negatively chirped\cite{note}. Depending on 

260: system parameters and harmonic order, one effect matches or 

261: dominates over the other. For instance, for the low-order 

262: harmonics in Fig.\ 4, the dynamic negative chirp is very small 

263: [Fig.\ 4(b)], and the plasma effect dominantly affects the 

264: time-frequency characteristics of high harmonics in such a way 

265: as to make high harmonics become positively chirped [Fig.\ 4(d)], 

266: leading to spectral line broadening [Fig.\ 4(c)]. In contrast, 

267: the dynamic negative chirp of higher-order harmonics in Fig.\ 5(b)

268: is substantially large, and the plasma-induced positive chirp 

269: more or less cancels out the dynamic chirp [Fig.\ 5(d)], resulting

270: in harmonic line narrowing [Fig.\ 5(c)]. In both cases of Figs.\ 4

271: and 5, the plasma effect blueshifts the central frequencies of 

272: high harmonics.

273: 

274: In conclusion, we have investigated the plasma-induced phase 

275: modulation of intense femtosecond laser pulses and its effects on

276: macroscopic high harmonic spectra, using a simple but accurate 1D

277: model. It has been shown that the plasmas produced in the medium 

278: induce a positive frequency chirp on the leading edge of the laser 

279: pulses. Strikingly, the plasma-induced chirp, which broadens the 

280: laser spectrum, can lead not only to broadening but also to 

281: narrowing of harmonic spectral lines. The underlying mechanism

282: has been explained by comparing the plasma-induced positive chirp 

283: with the dynamically-induced negative chirp. In stark contrast to 

284: the widely reported finding (in regimes well below saturation)

285: that high harmonics are negatively 

286: chirped\cite{Sali,Altu,Seki}, this study clearly 

287: demonstrates that, under laser fields of high intensity and gas 

288: pressures not too low, high harmonics can 

289: become free of chirp, or even become

290: positively chirped, thanks to the plasma effect. 

291: This suggests a way to control the chirp and spectral line shape 

292: of high harmonics using the plasma effect. 

293: 

294: This work was supported by the Ministry of Science and Technology 

295: of Korea through the Creative Research Initiative Program.

296: 

297: 

298: \begin{references}

299: \bibitem{Spiel} C. Spielmann, N. H. Burnett, S. Sartania, 

300:          R. Koppitsch, M. Schn\"{u}rer, C. Kan, M. Lenzner, 

301:          P. Wobrauschek, and K. C. Kulander, Science {\bf 278}, 

302:          661 (1997).

303: \bibitem{Chang} Z. Chang, A. Rundquist, H. Wang, M. M. Murnane, and

304:          H. C. Kapteyn, Phys. Rev. Lett {\bf 79}, 2967 (1997).

305: \bibitem{Kan1} C. Kan, N. H. Burnett, C. E. Capjack, and R. Rankin, 

306:          Phys. Rev. Lett. {\bf 79}, 2971 (1997).

307: \bibitem{Schn} M. Shn\"{u}rer, Z. Cheng, M. Hentschel, G. Tempea,

308:          P. K\'{a}lm\'{a}n, T. Brabec, and F. Krausz,

309:          Phys. Rev. Lett. {\bf 83}, 722 (1999).

310: \bibitem{Schaf} K. J. Schafer and K. C. Kulander, Phys. Rev. Lett.

311:          {\bf 78}, 638 (1997).

312: \bibitem{Lee} D. G. Lee, H. J. Shin, Y. H. Cha, K. H. Hong, J. H. Kim,

313:          and C. H. Nam, Phys. Rev. A {\bf 63}, 021801(R) (2001).

314: \bibitem{Watson} J. B. Watson, A. Sanpera, and K. Burnett,

315:          Phys. Rev. A {\bf 51}, 1458 (1995).

316: \bibitem{Kan2} C. Kan, C. E. Capjack, R. Rankin, and N. H. Burnett,

317:          Phys. Rev. A {\bf 52}, R4336 (1995).

318: \bibitem{Shin} H. J. Shin, D. G. Lee, Y. H. Cha, K. H. Hong, and 

319:          C. H. Nam, Phys. Rev. Lett. {\bf 83}, 2544 (1999);

320:          H. J. Shin, D. G. Lee, Y. H. Cha, J. H. Kim, K. H. Hong, and 

321:          C. H. Nam, Phys. Rev. A {\bf 63}, 053407 (2001).

322: \bibitem{Kim1} J. H. Kim, H. J. Shin, D. G. Lee, and C. H. Nam,

323:          Phys. Rev. A {\bf 62}, 055402 (2000). 

324: \bibitem{Gaarde} M. B. Gaarde, F. Salin, E. Constant, Ph. Balcou,

325:          K. J. Schafer, K. C. Kulander, and A. L'Huillier,

326:          Phys. Rev. A {\bf 59}, 1367 (1999).

327: \bibitem{Balcou} Ph. Balcou, A. S. Dederichs, M. B. Gaarde, and

328:          A. L'Huillier, J. Phys. B {\bf 32}, 2973 (1999).

329: \bibitem{Yablo} E. Yablonovitch, Phys. Rev. A {\bf 10}, 1888 (1974).

330: \bibitem{Wilks} S. C. Wilks, J. M. Dawson, and W. B. Mori, Phys. Rev.

331:          Lett. {\bf 61}, 337 (1988).

332: \bibitem{Wood} Wm. M. Wood, C. W. Siders, and M. C. Downer, Phys.

333:          Rev. Lett. {\bf 67}, 3523 (1991).

334: \bibitem{Rae} S. C. Rae and K. Burnett, Phys. Rev. A {\bf 46},

335:          1084 (1992).

336: \bibitem{Chessa} P. Chessa, E. De Wispelaere, F. Dorchies, 

337:          V. Malka, J. R. Marqu\`{e}s, G. Hamoniaux, P. Mora,

338:          and F. Amiranoff, Phys. Rev. Lett. {\bf 82}, 552 (1999).

339: \bibitem{Kim2} J. H. Kim, D. G. Lee, H. J. Shin, and C. H. Nam,

340:          Phys. Rev. A {\bf 63}, 063403 (2001). 

341: \bibitem{Ammo} M. V. Ammosov, N. B. Delone, and V. P. Krainov, 

342:          Sov. Phys. JETP {\bf 64}, 1191 (1986).

343: \bibitem{Sieg} A. E. Siegman, {\em Lasers} (Oxford University Press,

344:          Oxford, 1986). 

345: \bibitem{Sali} P. Sali\`{e}res, Ph. Antoine, A. de Bohan, and M.

346:          Lewenstein, Phys. Rev. Lett. {\bf 81}, 5544 (1998).

347: \bibitem{note} The laser pulse experienced by atoms in the case of 

348:          Fig.\ 3(b) resembles in form that used in our previous 

349:          discussions on the coherent sum method\cite{Kim1} and the

350:          time-frequency analysis\cite{Kim2}. Thus, the slope of 

351:          the Wigner distributions shown in Figs.\ 4(b) and 5(b)

352:          gives the dynamic chirp.

353: \bibitem{Altu} C. Altucci, C. Delfin, L. Roos, M. B. Gaarde, 

354:          A. L'Huillier, I. Mercer, T. Starczewski, and 

355:          C. G. Wahlstr\"{o}m, Phys. Rev. A {\bf 58}, 3934 (1998). 

356: \bibitem{Seki} T. Sekikawa, T. Ohno, T. Yamazaki, Y. Nabekawa, 

357:          and S. Watanabe, Phys. Rev. Lett. {\bf 83}, 2564 (1999).

358: \end{references}

359:  

360: 

361: \begin{figure}[h]

362: \centering

363: \epsfig{width=0.95 \linewidth,file=kim_fig1.ps}

364: \caption[ ]{

365: (a) Laser spectra obtained from Eq.\ (2) for a 30-fs (FWHM) laser pulse

366: of wavelength $\lambda =800$ nm after propagating through a Ne gas medium of

367: length $l=700$ $\mu$m.

368: The laser intensity and gas density are given by

369: $I=1\times 10^{15}$ W/cm$^2$ and $N_0=1\times 10^{18}$ cm$^{-3}$ ($\simeq$

370: 28 Torr) (thin solid line);

371: $I=1\times 10^{15}$ W/cm$^2$ and $N_0=3\times 10^{18}$ cm$^{-3}$

372: ($\simeq$ 85 Torr) (dotted line);

373: $I=3\times 10^{15}$ W/cm$^2$ and $N_0=1\times 10^{18}$ cm$^{-3}$

374: (dashed line). The spectrum of the incident laser pulse is drawn by

375: a thick solid line.

376: (b) Laser spectra obtained by direct numerical calculations of Eq.\ (1)

377: for the same parameters as in (a).}

378: \label{fig1}

379: \end{figure}

380: 

381: \begin{figure}[h]

382: \centering

383: \epsfig{width=0.95 \linewidth,file=kim_fig2.ps}

384: \caption[ ]{

385: Wigner distributions of the laser pulses specified in Fig.\ 1.

386: In (a)-(c) [(d)-(f)], the same parameters are used as for the thin solid,

387: dotted and dashed lines in Fig.\ 1(a) [Fig.\ 1(b)], respectively.

388: Only positive contour lines are shown.

389: }\label{fig2}

390: \end{figure}

391: 

392: \begin{figure}[h]

393: \centering

394: \epsfig{width=0.95 \linewidth,file=kim_fig3.ps}

395: \caption[ ]{

396: (a) Single-atom harmonic spectrum for a Ne atom placed at the

397: entrance of the medium.

398: (b) Macroscopic harmonic spectrum (in the far field, on the axis)

399: considering only the effect of laser focusing.

400: (c) Macroscopic harmonic spectrum taking into account both

401: the focusing and plasma effects.

402: A Ne gas medium of density $N_0=1\times 10^{18}$ cm$^{-3}$

403: and length $l=700$ $\mu$m is irradiated by a 30-fs (FWHM) Gaussian laser pulse

404: of wavelength $\lambda =800$ nm.

405: The medium is centered at $x=2$ mm behind the focus,

406: and because of focusing ($b=4$ mm) the laser intensity decreases

407: from $I=1\times 10^{15}$ W/cm$^2$ at the entrance

408: to $I=0.7\times 10^{15}$ W/cm$^2$ at the exit of the medium.

409: }\label{fig3}

410: \end{figure}

411: 

412: \begin{figure}[h]

413: \centering

414: \epsfig{width=0.95 \linewidth,file=kim_fig4.ps}

415: \caption[ ]{

416: (a) Enlarged view (on a linear scale) of Fig.\ 3(b) between $20\omega _0$

417: and $30\omega _0$ (only the focusing effect is considered).

418: (b) Wigner distribution of the harmonics chosen in (a).

419: (c) [(d)] Same as (a) [(b)] except that harmonics are

420: chosen from Fig.\ 3(c) (both the focusing and plasma effects are

421: considered). Positive and negative values of the Wigner distribution

422: are colored black and white, respectively.

423: }\label{fig4}

424: \end{figure}

425: 

426: \begin{figure}[h]

427: \centering

428: \epsfig{width=0.95 \linewidth,file=kim_fig5.ps}

429: \caption[ ]{

430: Same as Fig.\ 4 except that harmonics between $70\omega _0$ and $80\omega _0$

431: are chosen here.

432: }\label{fig5}

433: \end{figure}

434: 

435: 

436: \end{document}

437: 

438: