0808.2107/ms.tex
1: %\documentclass[12pt]{iopart}
2: 
3: \documentclass[twocolumn,preprintnumbers,amsmath,amssymb,prl,floatfix]{revtex4}
4: 
5: \usepackage{graphicx}
6: \usepackage{dcolumn}% Align table columns on decimal point
7: \usepackage{bm}% bold math
8: 
9: \newcommand{\eqb}{\begin{eqnarray}}
10: \newcommand{\eqe}{\end{eqnarray}}
11: \newcommand{\diff}{\textrm{d}}
12: \newcommand{\Eclassical}{E_{\rm cl}}
13: \newcommand{\Ecrit}{E_{\rm crit}}
14: \newcommand{\alphafine}{\alpha_{\rm f}}
15: \newcommand{\ehatvec}{\hat{{\bm e}}}
16: \newcommand{\betavec}{\bm{\beta}}
17: \newcommand{\lambdamicron}{\lambda_{\mu}}
18: \newcommand{\strength}{\hbox{$a$}}
19: \newcommand{\wsqcm}{\textrm{W\,cm}^{-2}}
20: \newcommand{\melec}{m_{\rm e}}
21: \begin{document} 
22: 
23: \title{Prolific pair production with high-power lasers}
24: 
25: \author{A.~R.~Bell}
26: \affiliation{Clarendon Laboratory, University of Oxford, 
27: Parks Road, Oxford OX1 3PU, UK} 
28: \affiliation{STFC Central Laser Facility, RAL, Didcot, OX11 0QX, UK}
29: \author{John~G.~Kirk}
30: \affiliation{Max-Planck-Institut f\"ur Kernphysik, Saupfercheckweg, 1, 
31: D-69117, Heidelberg, Germany}
32: \date{\today} 
33: 
34: \begin{abstract}
35:   Prolific electron-positron pair production is possible at laser intensities
36:   approaching 10$^{24}\,$Wcm$^{-2}$ at a wavelength of 1$\mu
37:   $m. An analysis of electron trajectories and interactions at the
38:   nodes ($B=0$) of two counter-propagating, circularly polarised laser
39:   beams shows that a cascade of $\gamma$-rays and pairs develops. The geometry
40:   is generalised qualitatively to linear polarisation and
41:   laser beams incident on a solid target.
42: \end{abstract}
43: 
44: \maketitle
45: 
46: High-power laser facilities have made dramatic progress recently, and 
47: the next few years may bring intensities of 
48: $10^{23}$--$10^{24}\,\wsqcm$ within reach. 
49: This naturally opens up new physics regimes \cite{kogaetal06,muelleretal08}.
50: The relativistic Lorentz
51: factor of an electron oscillating in vacuum in the electromagnetic field 
52: of 
53: a planar linearly polarised 
54: laser beam 
55: is 
56: $860\left(I_{24}\lambda _{\mu {\rm m}}^2\right)^{1/2}$
57: where $I_{24}$ is the laser 
58: intensity in $10^{24}\,\wsqcm$ and $\lambda_{\mu\rm m}$ is the laser
59: wavelength in micron. The corresponding peak
60: electric 
61: and magnetic fields are 
62: $2.7\times 10^{15}I_{24}^{1/2}\,$Vm$^{-1}$ 
63: and 
64: $91 I_{24}^{1/2}\,$GG. The  
65: Schwinger field $\Ecrit=1.3\times 10^{18}\,$Vm$^{-1}$
66: required for spontaneous electron-positron pair creation out of the vacuum 
67: would be attained
68: at a laser intensity of 
69: $2.3\times 10^{29}\,\wsqcm$ 
70: \cite{schwinger51,salaminetal06}.
71: Although this interesting regime 
72: is still far beyond projected laser intensities, 
73: several 
74: other strong-field
75: QED effects will soon be accessible to experiment.
76: In this {\em Letter\/} we show
77: how copious pair production by accelerated electrons interacting 
78: with the laser field can be achieved 
79: using laser
80: intensities $\sim10^{24}\,\wsqcm$. 
81: The key is to exploit the large 
82: transverse electromagnetic field seen by 
83: an electron when it experiences laser beams 
84: that are not propagating in parallel. We illustrate this effect
85: by computing the case of counter-propagating, circularly polarized beams.
86: The advantage offered by this configuration is analogous to the dramatic
87: increase in centre-of-mass energy when using colliding particle beams
88: instead of stationary targets. We argue that this advantage 
89: remains in less specific configurations such as 
90: tight focus and reflection from a solid surface.
91: Consequently, it may be possible to convert a large 
92: fraction of the laser energy into electron-positron pairs at a laser
93: intensity of $\sim10^{24}\,\wsqcm$ at approximately solid plasma 
94: density.
95: 
96: Relativistic electrons with Lorentz factor $\gamma$ 
97: moving perpendicular to a homogeneous 
98: magnetic field $B$ produce pairs if $\gamma B/B_{\rm crit}$
99: is greater than or of the order of unity, where,
100: $B_{\rm crit}=4.414\times10^4\,$GG is the magnetic equivalent of the 
101: Schwinger field $E_{\rm crit}$. The 
102: cross-sections for this and other relevant processes are well-known
103: \cite{erber66} and of interest also in astrophysics \cite{hardingli06}. 
104: Provided the electron trajectory can be approximated classically, 
105: these rates, when computed in a frame in which $\bm{E}\times\bm{B}=0$,  
106: are functions of the electric and magnetic fields only in the 
107: combination $E^2+B^2$ \cite{daughertylerche76,urrutia78}. 
108: Therefore, in a homogeneous electric field $\ehatvec E$, 
109: pair production occurs if the parameter
110: \eqb
111: \eta&=&\frac{\gamma E \sin \theta}{\Ecrit}
112: \label{etadef} 
113: \eqe  
114: is of order unity or larger, where 
115: $\theta$ is the angle between the electric
116: field and the electron momentum.
117: Pair
118: production by the trident process in 
119: which the electric field is provided by 
120: a high-Z nucleus and the Lorentz factor 
121: by accelerating electrons in the laser field,
122: has already been observed, but  
123: the process is relatively inefficient, and 
124: the yield achieved was $10^{-4}$ positrons for each fast
125: electron 
126: \cite{liangetal98,cowanetal99,nakashimaetal02}.  
127: Electron-positron
128: pairs have also been produced
129: by colliding $46.6\,$GeV electrons 
130: from a linear accelerator 
131: with an opposing laser beam, but this
132: produced a relatively modest number of pairs \cite{burkeetal97}.  
133: 
134: In a strong electromagnetic wave in vacuum, 
135: the Lorentz factor of an electron oscillates
136: about a value roughly equal to the strength parameter of the wave
137: \eqb
138: \strength&=&\frac{e E\lambda}{2\pi m c^2}
139: \nonumber
140: \\
141: &=&8.4\times 10^2 \left(I_{24}\lambda_{\mu{\rm m}}^2\right)^{1/2}
142: \label{strengthdef}
143: \eqe
144: \cite{lauetal03},
145: so that $\eta$ in Eq.~(\ref{etadef}) is approximately
146: $1.7 I_{24}\lambda_{\mu{\rm m}}$. This looks promising at first sight.
147: However, Eq.~(\ref{etadef}) assumes 
148: that $eE\sin\theta$ is the component of the particle's acceleration
149: perpendicular to its velocity, 
150: which is not the case in a laser field. In reality, 
151: an electron that is picked up by a single laser beam at initially low energy 
152: in the laboratory is accelerated on a trajectory that severely reduces the effective
153: value of $\eta$ below that in Eq.~(\ref{etadef}),
154: because the electric force is almost precisely cancelled
155: by that exerted by the magnetic field.
156: 
157: This can be understood in a way that brings
158: out the analogy with particle accelerators. In a 
159: plane electromagnetic wave in vacuum a charged particle has a
160: periodic trajectory in one special frame of reference 
161: (ignoring for
162: the moment radiation reaction) \cite{lauetal03}. 
163: This frame can be called the zero
164: momentum frame (ZMF).  
165: In it, all of 
166: the particle's phase-space variables are 
167: strictly periodic at the period of the wave, independently 
168: of its polarisation and waveform, and the particle energy 
169: oscillates around a value $\gamma mc^2\approx\strength mc^2$. 
170: In this frame, the electric and magnetic forces do not, in general,
171: cancel, 
172: and the perpendicular component of the acceleration is well-approximated by
173: $eE'$,  where $E'$ is the field
174: strength of the wave measured in the ZMF. Thus, 
175: the
176: importance of strong-field QED effects, such as pair-creation is
177: indeed determined by the parameter $\eta$, as defined in Eq.~(\ref{etadef}),
178: but computed in the ZMF, i.e., $\eta\approx \strength E'/\Ecrit$.  
179: In the case of a single laser beam
180: hitting a particle at rest, the ZMF does not coincide with the lab.\
181: frame, because the particle recoils. In fact, the ZMF moves in the
182: direction of propagation of the laser beam with a velocity
183: corresponding to a Lorentz factor equal to $\strength$ \cite{lauetal03}.  
184: The laser
185: frequency in the ZMF is thus red-shifted compared to the lab.\ frame. Because
186: the strength parameter $\strength$ is a Lorentz invariant, the reduced
187: wave frequency in the ZMF implies a reduced amplitude of the wave
188: field: $E'\approx E/\strength$. Consequently the (Lorentz invariant)
189: criterion for the importance of strong field QED effects becomes
190: $\strength E'/\Ecrit\approx E/\Ecrit>1$.  
191: In other words,
192: by using a single laser beam, the advantage gained over pure vacuum
193: effects by the relativistic oscillation of the electron is lost.
194: 
195: This is a consequence of the dual roles of accelerator and target
196: that are played by the laser beam. 
197: If the electron, instead of being initially at rest, is initially moving 
198: with a Lorentz factor $\gamma _{\rm init}$ that is much larger
199: than the strength parameter of the laser beam, as
200: in the experiment of Burke et al.\ \cite{burkeetal97}, the ZMF 
201: moves with almost the speed of the electron,
202: and $\eta\approx\gamma_{\rm init} E \sin \theta /\Ecrit$.
203: The initial Lorentz factor of the 
204: electron contributes to the threshold condition, but the
205: Lorentz factor due to oscillation in the laser field does not. In this case
206: the laser plays only the role of the target. 
207: 
208: However, as
209: in the case of intersecting particle beams, the situation can be
210: rescued if counter-propagating laser beams are employed. Then the
211: ZMF coincides with the lab.\ frame, and the importance of strong field
212: QED is again determined by Eq.~(\ref{etadef}).
213: A similar benefit is gained with a laser beam in tight focus --- which can
214: be decomposed into obliquely propagating plane waves ---
215: or with a beam in 
216: which a standing wave is set up when the laser encounters a dense
217: plasma.
218: Experimentally,
219: some of the 
220: most promising cases involve laser-solid interactions,
221: but the analysis of these is complicated.  Instead, we consider the 
222: theoretically simple case of pair production at the nodes 
223: ($B=0$) of two
224: counter-propagating circularly polarised laser beams of 
225: equal intensity.  The argument can then be qualitatively generalised
226: to laser-solid interactions.  
227: Strong-field QED effects  
228: at the nodes of 
229: counter-propagating waves have been considered previously 
230: \cite{brezinitzykson70,dipiazzaetal06},
231: but only for a vacuum in which the 
232: threshold condition relates to $E$ rather than $\gamma E$.
233: 
234: Classically, the  electron equation of motion, including radiation 
235: reaction according to the Landau \& Lifshitz
236: prescription
237: \cite{landaulifshitz75}, is
238: \eqb
239: \frac{\diff(\gamma \betavec)}{\diff t}&=&
240: -\frac{e}{{\melec c}}(\bm{E}+\betavec \times \bm{B})-
241: \nonumber\\
242: &&\frac{2 e^4\gamma^2}{3 m^3 c^5} 
243: \betavec \left|\bm{E}+\betavec \times \bm{B}\right|_\bot^2
244: \label{llequation}
245: \eqe 
246: The terms
247: that have been omitted here are of order $\gamma^{-2}$. 
248: The final term of eq.~(\ref{llequation}) represents the drag
249: and energy loss due to 
250: radiative emission, to which pair production is related, and is
251: proportional to the square of that component of the Lorentz force
252: $\bm{E}+\betavec \times \bm{B}$ perpendicular to $\betavec$.  In the case
253: of a planar uni-directional wave, the reduction in the electric field
254: in the ZMF is equivalent to the near cancellation of $\bm{E}$ with
255: $\bm{\beta} \times \bm{B}$ in the laboratory frame.
256: 
257: Two counter-propagating laser beams produce a standing wave with nodes
258: at which $B=0$ and the electric field rotates in direction 
259: with constant amplitude. 
260: By symmetry, an electron placed exactly at
261: the node does not move in the direction of the waves, but 
262: performs circular motion with the
263: centripetal force provided by the electric field. 
264: The equation of motion (\ref{llequation}) then 
265: simplifies to 
266: \eqb
267: \frac{\diff \left(\gamma\betavec\right)}{\diff t}&=&\frac{e E}{mc}\left\lbrace
268: -\ehatvec -\frac{2}{3} \betavec\gamma^2\frac{E}{\Eclassical}\left[
269: 1-\left(\betavec\cdot\ehatvec\right)^2\right]\right\rbrace
270: \label{eqmotionsimple}
271: \eqe
272: where $\hat{e}$ is the unit vector in the direction of the electric field
273: and we have defined the characteristic value $\Eclassical$ 
274: of the electric field in
275: classical electrodynamics: $\Eclassical=m^2c^4/e^3=\Ecrit/\alphafine$
276: ($\alphafine$ is the fine-structure constant).
277: As can be seen from Eq.~(\ref{eqmotionsimple}), 
278: the radiation reaction force becomes important 
279: when $\gamma^2 E/\Eclassical\sim 1$, i.e.,
280: $\eta\sim\left(\gamma\alphafine\right)^{-1}$,  
281: a situation that is reached 
282: at laser intensities $\sim10^{23}\lambdamicron^{-4/3}\,\wsqcm$.
283: The Landau \& Lifshitz prescription for the radiation reaction term 
284: is valid up to 
285: $\eta\approx 1/\alphafine$
286: \cite{dipiazza08}
287: but quantum effects already intervene at $\eta\sim 1$
288: \cite{erber66}.
289: 
290: Within a small fraction of a laser period, the electron trajectory
291: described by Eq.~(\ref{eqmotionsimple}) adjusts itself such that the
292: component of electric field parallel to $\betavec$ precisely
293: compensates the radiative losses, whilst the perpendicular component
294: enforces circular motion at the laser period, with Lorentz factor
295: $\gamma=\strength\sin \theta$. Combining these, and using the 
296: definition (\ref{etadef}) of the threshold parameter we can express
297: $E$ in terms of $\eta$.  
298: In an underdense plasma, the total electric field
299: $E$ is related to the intensity $I$ of each laser beam (separately)  
300: by $I=c E^2/(16\pi)$, which gives
301: \eqb
302: I_{24}&=&2.75\eta^4+0.28\left(\eta/\lambda_{\mu\rm m}\right)
303: \label{fluxeq}
304: \eqe
305: This relation is
306: plotted in Fig.~\ref{parameterspace} for a laser of wavelength
307: $1\,\mu $m.  In terms of these parameters, 
308: $\sin\theta=0.53\sqrt{\eta}\left(I_{24}\lambda_{\mu\rm m}\right)^{-1/2}$ --- 
309: at low intensity, the particle moves 
310: almost exactly perpendicularly to $\bm{E}$ 
311: and 
312: $\eta$ rises linearly with the laser intensity. However, when 
313: radiation reaction becomes important, this rise is slowed, and
314: $\eta=1$ is not achieved until $I_{24}=3$. The photons radiated 
315: because of the acceleration of the electron 
316: in the electric field of the laser --- which we term 
317: {\em curvature radiation}
318: \footnote{Frequently also called {\em bremsstrahlung}
319: \cite{erber66}, 
320: a term we reserve for radiation emitted in the electric field of a nucleus}
321: --- can
322: be described classically using the 
323: theory of synchrotron radiation. This predicts
324: that most radiated 
325: photons are emitted
326: with an energy 
327: \eqb
328: h\nu_{\rm s}&=&0.44\eta\gamma mc^2
329: \label{synchfreqdef}
330: \eqe where $\gamma mc^2=328\sqrt{\eta\lambda_{\mu{\rm m}}}\,$MeV is
331: the energy of the relativistic electron.  Because of quantum
332: effects analogous to the Klein-Nishina corrections to the Thomson
333: cross-section\cite{erber66,aharonian04}, 
334: the radiative energy loss does not proceed in the
335: continuous manner implied by Eq.~(\ref{eqmotionsimple}) when
336: $\eta>1$ and $I_{24}\gg1$. Neverthless, 
337: the classical trajectory is an adequate
338: approximation in the intensity range $10^{23}$--$10^{24}\wsqcm$, which
339: is of interest here, since the photon energy is significantly less
340: than the electron energy.
341: \begin{figure}
342: \includegraphics[width=0.5\textwidth]{fig1.ps}
343: \caption{%
344: \label{parameterspace}%
345: The parameters $\eta$ and $\chi$ --- see Eqs.~(\ref{etadef}) and
346: (\ref{chidef}) --- controlling the importance of electromagnetic
347: conversion by the accelerated electron and its curvature photon
348: respectively, as a function of laser intensity, for a laser wavelength
349: of $1\mu\textrm{m}$. Also plotted are the optical depth
350: $\tau$ of the curvature  
351: photon across one
352: wavelength, the number $N_\pm$
353: of pairs produced per electron in one laser
354: period by both this process and by that of
355: pair production by the trident process --- labelled $\tau_{\rm tr}$,
356: and the number $N_{\gamma}$
357: of curvature radiation photons produced per electron
358: per laser period.}
359: \end{figure}
360: 
361: However, other important quantum effects are already present 
362: at intensities in this range. There are two 
363: processes that produce electron-positron pairs.  At low laser
364: intensities, the trident process dominates, 
365: in which an electron produces an
366: electron-positron pair via an intermediate virtual photon. 
367: In a homogeneous electric or magnetic field (a good
368: approximation when $\lambda_{\rm laser}\gg
369: h/mc=2.4\times10^{-6}\mu\textrm{m}$) the rate is given by
370: \cite{erber66,urrutia78}. Expressed as a production rate per electron
371: per laser period, it can be written $\tau_{\rm tr}=0.06\left
372:   (I_{24}\lambda_{\mu{\rm m}}^2\right
373: )^{1/2}\eta^{1/4}\textrm{exp}\left(-8/\sqrt{3\eta}\right)$ for
374: $\eta\ll1$, and, for $\eta>1$, it goes over to a slow logarithmic
375: increase. The precise form is plotted in Fig.~\ref{parameterspace}.
376: 
377: At higher intensities, the related process becomes important, 
378: in which the electron first
379: produces a real photon by curvature radiation, which subsequently
380: creates a pair.  However, to compare this to the
381: trident process one must specify the distance over which the real
382: photon is permitted to undergo conversion: if this is large, all
383: photons will convert into pairs, whereas if it is very short, none of
384: them will. The two processes are almost equal in rate if, at
385: $\eta\approx1$, the distance is chosen to be
386: $(\hbar/mc)(\Ecrit/E)=1.3\times10^{-4}I_{24}^{-1/2}\mu\textrm{m}$
387: \cite{erber66}.
388: However, in reality, this length is determined by the size of the region in 
389: which the laser beams overlap, which we 
390: conservatively assume to be $\lambda_{\rm laser}$.
391: This gives the process that involves a real photon as intermediary 
392: a substantial advantage. An additional, though less important, advantage arises 
393: because the propagation direction of 
394: the photon does not rotate, and, therefore, 
395: the perpendicular component of electric field it experiences is 
396: $\sim E$, rather than $ E \sin \theta$. 
397: The absorption coefficient is controlled by the parameter
398: \eqb
399: \chi&=&\frac{h\nu_{\rm s}E}{2mc^2\Ecrit} 
400: \label{chidef}
401: \eqe 
402: From Eq.~(\ref{synchfreqdef}) and writing $E$ in terms of $I_{24}$,
403: $\chi=0.42\eta^{3/2}\sqrt{I_{24}\lambda_{\mu{\rm m}}}$ and this 
404: function is plotted in Fig.~\ref{parameterspace}.  The 
405: photon optical depth to absorption in a path length $\lambda_{\rm laser}$ is
406: $\tau=12.8
407: \left(I_{24}\lambda_{\mu{\rm m}}^2\right)
408: \textrm{exp}\left[-4/\left(3\chi\right)\right]$, for
409: $\chi\ll1$, peaking at $\chi\approx8$ and falling off for larger
410: $\chi$ \cite{erber66}. It is also plotted in
411: Fig.~\ref{parameterspace}.
412: The total pair-production rate per electron
413: per laser period is the product of the photon absorption probability 
414: $1-\textrm{exp}\left(-\tau\right)$ in a length
415: $\lambda_{\rm laser}$ multiplied by the rate of production of photons by
416: curvature radiation. This quantity, together with the number of curvature
417: radiation photons 
418: emitted per
419: electron per laser period
420: (i.e., the energy radiated divided by $h\nu_{\rm s}$): 
421: $N_{\gamma}=6.42\alpha_{\rm f}\gamma$ is also 
422: shown in Fig.~\ref{parameterspace}.
423: 
424: Inspection of this figure shows that for laser intensities less than
425: roughly $I=3.3\times 10^{23}\,\wsqcm$, where $\eta=0.51$, pair
426: production is dominated by the trident process. At this intensity, each
427: electron in the zone where the laser beams overlap produces on average
428: $3\times 10^{-5}$ pairs in a single laser period. The curvature
429: radiation energy losses, which are $123\,$kW per electron, dominate
430: over pair production.  They are sufficient to damp the laser beams in
431: $1.8n_{23}^{-1}\,$fsec where $n_{23}$ is the electron density in
432: $10^{23}\,$cm$^{-3}$.  The total number of pairs produced 
433: in the absence of other energy losses is $7\times 10^4$
434: per Joule of laser energy.  
435: 
436: At
437: intensities above $I=3.3\times10^{23}\,\wsqcm$, the number of pairs
438: produced by photon-induced pair production rises steeply. 
439: These pairs are 
440: also accelerated and generate additional photons and pairs.  A cascade
441: should develop when $N_{\pm}\approx 1$, which occurs at 
442: $I_{24}\approx 1$ and $\eta\approx 0.7$.  
443: At this intensity, the laser power should be
444: divided roughly equally between photons and pairs with energy
445: $\sim 80\,$MeV per photon and per pair. This process is not sensitive to 
446: the number of electrons
447: initially in the interaction region. Complete conversion of laser energy
448: to photons and pairs implies the production of $\sim4\times 10^{10}$ pairs
449: per Joule of laser energy.  The precise conditions under which
450: a cascade is initiated are, however, sensitive to 
451: geometrical effects related to the intersection angle and the
452: intersection volume of the laser beams.
453: 
454: For simplicity of analysis we have assessed pair-production at the
455: nodes of counter-propagating circularly polarised waves and found that
456: the condition for pair production is 
457: $\strength E_\perp>\Ecrit$
458: where 
459: $\strength$ is the strength parameter, roughly equal to the 
460: Lorentz factor of an
461: electron oscillating in the electromagnetic wave.  This contrasts with
462: the much higher threshold in the case of an electron overtaken by a
463: planar uni-directional wave.  
464: The uni-directional plane wave is a special case, and the conditions
465: for prolific pair-production should be met at laser intensities $\sim
466: 10 ^{24}\,\wsqcm$ away from the node or when the polarisation is not
467: circular, although the analysis is more complicated and numerical
468: factors of order unity will change the exact quantitative
469: result. Conditions for pair production may also occur at similar laser
470: intensities for a single laser beam incident on an overdense solid
471: target. The incident and reflected laser beams form
472: counter-propagating waves.  Even if the laser beam is substantially
473: absorbed rather than reflected, the electric field swells as the wave
474: passes into the plasma and the phase velocity differs from the speed
475: of light so that $\bm{E}$ and $\betavec\times\bm{B}$ 
476: are unlikely to cancel as in
477: the uni-directional case. Furthermore, if the laser beam is in tight
478: focus, as required for the highest intensities, it can be decomposed
479: into obliquely propagating plane waves. 
480: Another factor favouring pair production is
481: that the electrostatic field required for quasi-neutrality at a solid
482: surface stops the electrons moving freely in the direction of laser
483: propagation so that $\bm{E}$ and $\betavec\times\bm{B}$ 
484: once again cannot cancel, and the laser is  
485: able to play the 
486: role of particle accelerator and target simultaneously.
487: 
488: We predict that pair-production should be a standard feature of
489: laser-plasma interactions at intensities in the range $3 \times 10
490: ^{23}-10 ^{24}$Wcm$^{-2}$ at a laser wavelength of 1$\mu $m. 
491: A significant number of pairs is produced at the lower end of this
492: intensity range and the number increases dramatically when the
493: intensity approaches $10 ^{24}\,\wsqcm$.  At intensities $\sim 10
494: ^{24}\,\wsqcm$ a cascade sets in, producing an avalanche which
495: efficiently converts the laser energy into 
496: roughly equal numbers of pairs and $\gamma$-rays.
497: These predictions may be
498: tested using high-power lasers in the next few years.
499: 
500: \section*{Acknowledgments}
501: This research was carried out at 
502: the St John's College Research Centre, University of Oxford, and we
503: thank the members of the Centre for their warm hospitality. 
504: \begin{thebibliography}{18}
505: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
506: \expandafter\ifx\csname bibnamefont\endcsname\relax
507:   \def\bibnamefont#1{#1}\fi
508: \expandafter\ifx\csname bibfnamefont\endcsname\relax
509:   \def\bibfnamefont#1{#1}\fi
510: \expandafter\ifx\csname citenamefont\endcsname\relax
511:   \def\citenamefont#1{#1}\fi
512: \expandafter\ifx\csname url\endcsname\relax
513:   \def\url#1{\texttt{#1}}\fi
514: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
515: \providecommand{\bibinfo}[2]{#2}
516: \providecommand{\eprint}[2][]{\url{#2}}
517: 
518: \bibitem[{\citenamefont{{Koga} et~al.}(2006)\citenamefont{{Koga}, {Esirkepov},
519:   and {Bulanov}}}]{kogaetal06}
520: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{{Koga}}},
521:   \bibinfo{author}{\bibfnamefont{T.~Z.} \bibnamefont{{Esirkepov}}},
522:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{S.~V.}
523:   \bibnamefont{{Bulanov}}}, \bibinfo{journal}{Journal of Plasma Physics}
524:   \textbf{\bibinfo{volume}{72}}, \bibinfo{pages}{1315} (\bibinfo{year}{2006}).
525: 
526: \bibitem[{\citenamefont{{M{\"u}ller} et~al.}(2008)\citenamefont{{M{\"u}ller},
527:   {di Piazza}, {Shahbaz}, {B{\"u}rvenich}, {Evers}, {Hatsagortsyan}, and
528:   {Keitel}}}]{muelleretal08}
529: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{{M{\"u}ller}}},
530:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{{di Piazza}}},
531:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{{Shahbaz}}},
532:   \bibinfo{author}{\bibfnamefont{T.~J.} \bibnamefont{{B{\"u}rvenich}}},
533:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{{Evers}}},
534:   \bibinfo{author}{\bibfnamefont{K.~Z.} \bibnamefont{{Hatsagortsyan}}},
535:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{C.~H.}
536:   \bibnamefont{{Keitel}}}, \bibinfo{journal}{Laser Physics}
537:   \textbf{\bibinfo{volume}{18}}, \bibinfo{pages}{175} (\bibinfo{year}{2008}).
538: 
539: \bibitem[{\citenamefont{{Schwinger}}(1951)}]{schwinger51}
540: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{{Schwinger}}},
541:   \bibinfo{journal}{Physical Review} \textbf{\bibinfo{volume}{82}},
542:   \bibinfo{pages}{664} (\bibinfo{year}{1951}).
543: 
544: \bibitem[{\citenamefont{{Salamin} et~al.}(2006)\citenamefont{{Salamin}, {Hu},
545:   {Hatsagortsyan}, and {Keitel}}}]{salaminetal06}
546: \bibinfo{author}{\bibfnamefont{Y.~I.} \bibnamefont{{Salamin}}},
547:   \bibinfo{author}{\bibfnamefont{S.~X.} \bibnamefont{{Hu}}},
548:   \bibinfo{author}{\bibfnamefont{K.~Z.} \bibnamefont{{Hatsagortsyan}}},
549:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{C.~H.}
550:   \bibnamefont{{Keitel}}}, \bibinfo{journal}{Physics Reports}
551:   \textbf{\bibinfo{volume}{427}}, \bibinfo{pages}{41} (\bibinfo{year}{2006}).
552: 
553: \bibitem[{\citenamefont{{Erber}}(1966)}]{erber66}
554: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{{Erber}}},
555:   \bibinfo{journal}{Reviews of Modern Physics} \textbf{\bibinfo{volume}{38}},
556:   \bibinfo{pages}{626} (\bibinfo{year}{1966}).
557: 
558: \bibitem[{\citenamefont{{Harding} and {Lai}}(2006)}]{hardingli06}
559: \bibinfo{author}{\bibfnamefont{A.~K.} \bibnamefont{{Harding}}}
560:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{{Lai}}},
561:   \bibinfo{journal}{Reports of Progress in Physics}
562:   \textbf{\bibinfo{volume}{69}}, \bibinfo{pages}{2631} (\bibinfo{year}{2006}),
563:   \eprint{arXiv:astro-ph/0606674}.
564: 
565: \bibitem[{\citenamefont{{Daugherty} and {Lerche}}(1976)}]{daughertylerche76}
566: \bibinfo{author}{\bibfnamefont{J.~K.} \bibnamefont{{Daugherty}}}
567:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{{Lerche}}},
568:   \bibinfo{journal}{\prd} \textbf{\bibinfo{volume}{14}}, \bibinfo{pages}{340}
569:   (\bibinfo{year}{1976}).
570: 
571: \bibitem[{\citenamefont{{Urrutia}}(1978)}]{urrutia78}
572: \bibinfo{author}{\bibfnamefont{L.~F.} \bibnamefont{{Urrutia}}},
573:   \bibinfo{journal}{\prd} \textbf{\bibinfo{volume}{17}}, \bibinfo{pages}{1977}
574:   (\bibinfo{year}{1978}).
575: 
576: \bibitem[{\citenamefont{{Liang} et~al.}(1998)\citenamefont{{Liang}, {Wilks},
577:   and {Tabak}}}]{liangetal98}
578: \bibinfo{author}{\bibfnamefont{E.~P.} \bibnamefont{{Liang}}},
579:   \bibinfo{author}{\bibfnamefont{S.~C.} \bibnamefont{{Wilks}}},
580:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{{Tabak}}},
581:   \bibinfo{journal}{Physical Review Letters} \textbf{\bibinfo{volume}{81}},
582:   \bibinfo{pages}{4887} (\bibinfo{year}{1998}).
583: 
584: \bibitem[{\citenamefont{{Cowan} et~al.}(1999)\citenamefont{{Cowan}, {Perry},
585:   {Key}, {Ditmire}, {Hatchett}, {Henry}, {Moody}, {Moran}, {Pennington},
586:   {Phillips} et~al.}}]{cowanetal99}
587: \bibinfo{author}{\bibfnamefont{T.~E.} \bibnamefont{{Cowan}}},
588:   \bibinfo{author}{\bibfnamefont{M.~D.} \bibnamefont{{Perry}}},
589:   \bibinfo{author}{\bibfnamefont{M.~H.} \bibnamefont{{Key}}},
590:   \bibinfo{author}{\bibfnamefont{T.~R.} \bibnamefont{{Ditmire}}},
591:   \bibinfo{author}{\bibfnamefont{S.~P.} \bibnamefont{{Hatchett}}},
592:   \bibinfo{author}{\bibfnamefont{E.~A.} \bibnamefont{{Henry}}},
593:   \bibinfo{author}{\bibfnamefont{J.~D.} \bibnamefont{{Moody}}},
594:   \bibinfo{author}{\bibfnamefont{M.~J.} \bibnamefont{{Moran}}},
595:   \bibinfo{author}{\bibfnamefont{D.~M.} \bibnamefont{{Pennington}}},
596:   \bibinfo{author}{\bibfnamefont{T.~W.} \bibnamefont{{Phillips}}},
597:   \bibnamefont{et~al.}, \bibinfo{journal}{Laser Part.\ Beams}
598:   \textbf{\bibinfo{volume}{17}}, \bibinfo{pages}{773} (\bibinfo{year}{1999}).
599: 
600: \bibitem[{\citenamefont{{Nakashima} et~al.}(2002)\citenamefont{{Nakashima},
601:   {Cowan}, and {Takabe}}}]{nakashimaetal02}
602: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{{Nakashima}}},
603:   \bibinfo{author}{\bibfnamefont{T.~E.} \bibnamefont{{Cowan}}},
604:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{{Takabe}}},
605:   in \emph{\bibinfo{booktitle}{Science of Superstrong Field Interactions}},
606:   edited by \bibinfo{editor}{\bibfnamefont{K.}~\bibnamefont{{Nakajima}}}
607:   \bibnamefont{and} \bibinfo{editor}{\bibfnamefont{M.}~\bibnamefont{{Deguchi}}}
608:   (\bibinfo{year}{2002}), vol. \bibinfo{volume}{634} of
609:   \emph{\bibinfo{series}{American Institute of Physics Conference Series}}, pp.
610:   \bibinfo{pages}{323--328}.
611: 
612: \bibitem[{\citenamefont{{Burke} et~al.}(1997)\citenamefont{{Burke}, {Field},
613:   {Horton-Smith}, {Spencer}, {Walz}, {Berridge}, {Bugg}, {Shmakov},
614:   {Weidemann}, {Bula} et~al.}}]{burkeetal97}
615: \bibinfo{author}{\bibfnamefont{D.~L.} \bibnamefont{{Burke}}},
616:   \bibinfo{author}{\bibfnamefont{R.~C.} \bibnamefont{{Field}}},
617:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{{Horton-Smith}}},
618:   \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{{Spencer}}},
619:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{{Walz}}},
620:   \bibinfo{author}{\bibfnamefont{S.~C.} \bibnamefont{{Berridge}}},
621:   \bibinfo{author}{\bibfnamefont{W.~M.} \bibnamefont{{Bugg}}},
622:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{{Shmakov}}},
623:   \bibinfo{author}{\bibfnamefont{A.~W.} \bibnamefont{{Weidemann}}},
624:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{{Bula}}},
625:   \bibnamefont{et~al.}, \bibinfo{journal}{Physical Review Letters}
626:   \textbf{\bibinfo{volume}{79}}, \bibinfo{pages}{1626} (\bibinfo{year}{1997}).
627: 
628: \bibitem[{\citenamefont{{Lau} et~al.}(2003)\citenamefont{{Lau}, {He},
629:   {Umstadter}, and {Kowalczyk}}}]{lauetal03}
630: \bibinfo{author}{\bibfnamefont{Y.~Y.} \bibnamefont{{Lau}}},
631:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{{He}}},
632:   \bibinfo{author}{\bibfnamefont{D.~P.} \bibnamefont{{Umstadter}}},
633:   \bibnamefont{and}
634:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{{Kowalczyk}}},
635:   \bibinfo{journal}{Physics of Plasmas} \textbf{\bibinfo{volume}{10}},
636:   \bibinfo{pages}{2155} (\bibinfo{year}{2003}).
637: 
638: \bibitem[{\citenamefont{{Brezin} and {Itzykson}}(1970)}]{brezinitzykson70}
639: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{{Brezin}}} \bibnamefont{and}
640:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{{Itzykson}}},
641:   \bibinfo{journal}{\prd} \textbf{\bibinfo{volume}{2}}, \bibinfo{pages}{1191}
642:   (\bibinfo{year}{1970}).
643: 
644: \bibitem[{\citenamefont{{di Piazza} et~al.}(2006)\citenamefont{{di Piazza},
645:   {Hatsagortsyan}, and {Keitel}}}]{dipiazzaetal06}
646: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{{di Piazza}}},
647:   \bibinfo{author}{\bibfnamefont{K.~Z.} \bibnamefont{{Hatsagortsyan}}},
648:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{C.~H.}
649:   \bibnamefont{{Keitel}}}, \bibinfo{journal}{Physical Review Letters}
650:   \textbf{\bibinfo{volume}{97}}, \bibinfo{pages}{083603}
651:   (\bibinfo{year}{2006}), \eprint{arXiv:hep-ph/0602039}.
652: 
653: \bibitem[{\citenamefont{{Landau} and {Lifshitz}}(1975)}]{landaulifshitz75}
654: \bibinfo{author}{\bibfnamefont{L.~D.} \bibnamefont{{Landau}}} \bibnamefont{and}
655:   \bibinfo{author}{\bibfnamefont{E.~M.} \bibnamefont{{Lifshitz}}},
656:   \emph{\bibinfo{title}{{The classical theory of fields}}}
657:   (\bibinfo{publisher}{Course of theoretical physics - Pergamon International
658:   Library of Science, Technology, Engineering and Social Studies, Oxford:
659:   Pergamon Press, 1975, 4th rev.engl.ed.}, \bibinfo{year}{1975}).
660: 
661: \bibitem[{\citenamefont{{di~Piazza}}(2008)}]{dipiazza08}
662: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{{di~Piazza}}},
663:   \bibinfo{journal}{Lett.\ Math.\ Phys.} \textbf{\bibinfo{volume}{83}},
664:   \bibinfo{pages}{305} (\bibinfo{year}{2008}), \eprint{arXiv:0801.1751}.
665: 
666: \bibitem[{\citenamefont{{Aharonian}}(2004)}]{aharonian04}
667: \bibinfo{author}{\bibfnamefont{F.~A.} \bibnamefont{{Aharonian}}},
668:   \emph{\bibinfo{title}{{Very high energy cosmic gamma radiation : a crucial
669:   window on the extreme Universe}}} (\bibinfo{publisher}{Very high energy
670:   cosmic gamma radiation : a crucial window on the extreme Universe, by
671:   F.A.~Aharonian.~River Edge, NJ: World Scientific Publishing, 2004},
672:   \bibinfo{year}{2004}).
673: 
674: \end{thebibliography}
675: \end{document}
676: