1: \documentclass[12pt]{article}
2: \textheight=22cm
3: \textwidth=14.4cm
4: \topmargin=-1.5cm
5: \leftmargin=0.5cm
6: \overfullrule=0pt
7: \def\int {\intop \limits}
8: \def\oint {\ointop \limits}
9: \def\fnote#1{\footnote}
10: \begin{document}
11:
12: \renewcommand \theequation{\thesection.\arabic{equation}}
13:
14: \title{Radiation from polarized electrons in oriented crystals
15: at high energy}
16: \author{V. N. Baier
17: and V. M. Katkov\\
18: Budker Institute of Nuclear Physics\\ 630090 Novosibirsk,
19: Russia}
20:
21: \maketitle
22:
23: \begin{abstract}
24: Radiation from high energy electrons in an oriented crystal can be
25: considered in a frame of the quasiclassical operator method
26: which appears to be a most satisfactory approach to the problem.
27: Under some quite generic assumptions the general expression is derived
28: for the probability of circularly polarized photon emission
29: from the longitudinally polarized electron in oriented crystal.
30: The particular mechanism of radiation depends
31: on interrelation between the angle of incidence $\vartheta_0$ (angle
32: between the momentum of initial electron and axis (plane) of crystal) and
33: angle $\vartheta_v \equiv V_0/m$~($V_0$ is the scale of a potential
34: of axis or a plane relative to which the angle $\vartheta_0$ is defined).
35: When $\vartheta_0 \ll \vartheta_v$
36: one has magnetic bremsstrahlung type of radiation (with corrections
37: $\propto \vartheta_0^2$ which are due to inhomogeneous character of field in
38: crystal). When $\vartheta_0 \gg \vartheta_v$ one obtains the theory of
39: coherent bremsstrahlung, while for $\vartheta_0 \geq \vartheta_v$ one
40: arrives to the modified theory of coherent bremsstrahlung. At high energy
41: radiation in oriented crystals is strongly enhanced comparing with standard
42: bremsstrahlung.
43:
44:
45: \end{abstract}
46:
47: \newpage
48:
49: \section{Introduction}
50:
51: The quasiclassical operator method developed by
52: authors \cite{BK0}-\cite{BK2} is adequate for consideration of the
53: electromagnetic processes at high energy.
54: The probability of photon emission has a form (see
55: \cite{BKS}, p.63, Eq.(2.27); the method is given also in
56: \cite{BLP},\cite{BKF})
57: \begin{equation}
58: dw=\frac{e^2}{(2\pi)^2} \frac{d^3k}{\omega}
59: \int_{}^{}dt_2\int_{}^{}dt_1 R^{\ast}(t_2) R(t_1)
60: \exp \left[-\frac{i\varepsilon}{\varepsilon'}
61: \left(kx(t_2)-kx(t_1)\right) \right],
62: \label{1}
63: \end{equation}
64: where $k^{\mu}=(\omega, {\bf k})$ is the 4-momentum of the emitted
65: photon, $k^2=0$, $x^{\mu}(t)=(t, {\bf r}(t))$, $t$ is the time,
66: and ${\bf r}(t)$ is the particle location on a classical trajectory,
67: $kx(t)=\omega t- {\bf kr}(t)$,
68: $\varepsilon'=\varepsilon-\omega$,
69: we employ units such that $\hbar=c=1$.
70: The matrix element $R(t)$ is defined by the structure of a current.
71: For an electron (spin 1/2 particle) one has
72: \begin{eqnarray}
73: && R(t)=\frac{m}{\sqrt{\varepsilon \varepsilon'}}
74: \overline{u}_{s_{f}}({\bf p'})\hat{e}^{\ast}u_{s_{i}}({\bf p})
75: =\varphi_{s_{f}}^{+}\left(A(t)+i\mbox{\boldmath$\sigma$}{\bf B}(t) \right)
76: \varphi_{s_{i}},
77: \nonumber \\
78: && A(t)= \frac{1}{2}\left(1+\frac{\varepsilon}{\varepsilon'} \right)
79: {\bf e}^{\ast}\mbox{\boldmath$\vartheta$}(t),
80: \nonumber \\
81: && {\bf B(t)}=\frac{\omega}{2\varepsilon'}\left({\bf e}^{\ast} \times
82: \left(\frac{{\bf n}}{\gamma}- \mbox{\boldmath$\vartheta$}(t)\right) \right),
83: \label{2}
84: \end{eqnarray}
85: here ${\bf e}$ is the vector of the polarization of a photon
86: (the Coulomb gauge is used),the four-component spinors $u_{s_f},
87: u_{s_i}$ and the two-component spinors $\varphi_{s_f},
88: \varphi_{s_i}$ describe the initial ($s_i$) and final ($s_f$)
89: polarization of the electron, {\bf v}={\bf v}(t) is the electron
90: velocity, $\mbox{\boldmath$\vartheta$}(t) = \left({\bf v}-{\bf
91: n}\right) \simeq {\bf v}_{\perp}(t)$, ${\bf
92: v}_{\perp}$ is the component of particle velocity perpendicular
93: to the vector ${\bf n}={\bf k}/|{\bf k}|$, $\gamma=\varepsilon/m$
94: is the Lorentz factor. The expressions in Eq.(\ref{2}) are
95: given for radiation of ultrarelativistic electrons, they are
96: written down with relativistic accuracy (terms $\sim 1/\gamma$
97: are neglected) and in the small angle approximation.
98:
99: The important parameter $\chi$
100: characterizes the quantum effects in an external field, when
101: $\chi \ll 1$ we are in the classical domain and with $\chi \geq
102: 1$ we are already well inside the quantum domain
103: \begin{equation}
104: \chi =\frac{|{\bf F}|\gamma}{F_0}, \quad {\bf F}={\bf
105: E}_{\perp}+({\bf v}\times {\bf H}),\quad {\bf E}_{\perp}={\bf E}-
106: {\bf v}({\bf v}{\bf E}),
107: \label{3}
108: \end{equation}
109: where ${\bf E} ({\bf H})$ is an electric (magnetic) field,
110: $F_0=m^2/e=(m^2c^2/e\hbar)$ is the quantum boundary (Schwinger)
111: field: $H_0=4.41 \cdot 10^{13} {\rm Oe},~ E_0=1.32\cdot 10^{16}
112: {\rm V/cm}$.
113:
114: The quasiclassical operator method is applicable when $H \ll H_0,~
115: E \ll E_0$ and $\gamma \gg 1$.
116:
117: The general expression for combination $R^{\ast}(t_2) R(t_1)=R_2^{\ast}R_1$
118: with all polarization taken into account is
119: \begin{equation}
120: R_2^{\ast}R_1 \equiv N_{21}(\mbox{\boldmath$\zeta$}_i,
121: \mbox{\boldmath$\zeta$}_f, \textbf{e})=\frac{1}{4}
122: {\rm Tr}\left[(1+\mbox{\boldmath$\zeta$}_i \mbox{\boldmath$\sigma$})
123: \left(A_2^{\ast}-i \mbox{\boldmath$\sigma$}\textbf{B}_2^{\ast} \right)
124: (1+\mbox{\boldmath$\zeta$}_f \mbox{\boldmath$\sigma$})
125: \left(A_1+i \mbox{\boldmath$\sigma$}\textbf{B}_1 \right)\right],
126: \label{4}
127: \end{equation}
128: where we use for the description of electron polarization
129: the vector $\mbox{\boldmath$\zeta$}$ describing the
130: polarization of the electron (in its rest frame),
131: $\mbox{\boldmath$\zeta$}_i$ is the spin vector of initial electron,
132: $\mbox{\boldmath$\zeta$}_f$ is the spin vector of final electron.
133: Summing over final spin states we have
134: \begin{equation}
135: \sum_{\zeta_f}R_2^{\ast}R_1 =A_2^{\ast}A_1
136: +\textbf{B}_2^{\ast}\textbf{B}_1
137: +i\left[A_2^{\ast}(\mbox{\boldmath$\zeta$}_i\textbf{B}_1)
138: -A_1 (\mbox{\boldmath$\zeta$}_i\textbf{B}_2^{\ast})
139: +\mbox{\boldmath$\zeta$}_i\left(\textbf{B}_2^{\ast}\times
140: \textbf{B}_1\right) \right],
141: \label{5}
142: \end{equation}
143: where the two first terms describe the radiation of unpolarized
144: electrons and the last terms is an addition dependent on
145: the initial spin.
146:
147: A combination of matrix elements $R_2^{\ast}R_1$ for spin-flip
148: transition follows from Eq.(\ref{4}) after substitution
149: $\mbox{\boldmath$\zeta$}_i =\mbox{\boldmath$\zeta$}$,
150: $\mbox{\boldmath$\zeta$}_f=-\mbox{\boldmath$\zeta$}$
151: \begin{equation}
152: N_{21}(\mbox{\boldmath$\zeta$}, -\mbox{\boldmath$\zeta$}, \textbf{e})=
153: \textbf{B}_2^{\ast}\textbf{B}_1-(\mbox{\boldmath$\zeta$}\textbf{B}_2^{\ast})
154: (\mbox{\boldmath$\zeta$}\textbf{B}_1)-i\mbox{\boldmath$\zeta$}
155: \left(\textbf{B}_2^{\ast}\times\textbf{B}_1\right).
156: \label{6}
157: \end{equation}
158: It depends on the spin-flip amplitudes $\textbf{B}_{1,2}$ only.
159:
160: After summing in Eq.(\ref{5}) over the polarization of emitted photon
161: $\lambda$ we obtain for unpolarized electrons
162: \begin{equation}
163: \sum_{\lambda, \zeta_f} R_2^{\ast} R_1 \rightarrow \frac{1}{2\varepsilon'^2}
164: \left[\frac{\omega^2}{\gamma^2}+\left(\varepsilon^2+\varepsilon'^2 \right)
165: \mbox{\boldmath$\vartheta$}_1\mbox{\boldmath$\vartheta$}_2 \right].
166: \label{7}
167: \end{equation}
168: For the longitudinally polarized initial electron and for circular
169: polarization of emitted photon we have
170: \begin{equation}
171: \sum_{\lambda, \zeta_f} R_2^{\ast}R_1 = \frac{1}{2\varepsilon'^2}
172: \left\{\frac{\omega^2}{\gamma^2}(1+\xi)+
173: \left[(1+\xi)\varepsilon^2+(1-\xi)\varepsilon'^2
174: \right]\mbox{\boldmath$\vartheta$}_1\mbox{\boldmath$\vartheta$}_2 \right\},
175: \label{8}
176: \end{equation}
177: where $\xi=\lambda\zeta,~\lambda=\pm 1$ is the helicity of emitted photon,
178: $\zeta =\pm 1$ is the helicity of the initial electron.
179: In this expression we omit the terms which vanish after integration
180: over angles of emitted photon.
181:
182: Just for selected polarizations there are the strong effects
183: in the radiation probability in high-energy limit: in the hard part
184: of spectrum the longitudinally polarized initial electron emits
185: mainly circularly polarized photon. This is a particular example of
186: helicity transfer.
187:
188: \section{General approach to radiation in oriented crystals}
189: \setcounter{equation}{0}
190:
191: The theory of high-energy electron radiation and electron-positron pair
192: creation in oriented crystals was developed in \cite{BKS1}-\cite{BKS2},
193: and given in \cite{BKS}. In these publications the radiation
194: from unpolarized electrons was considered including
195: the polarization density matrix
196: of emitted photons. Since Eq.(\ref{8}) has the same structure as
197: Eq.(\ref{7}), below we use systematically the methods of mentioned papers
198: to obtain the characteristics of radiation
199: from longitudinally polarized electron.
200:
201: Let us remind that along with the parameter $\chi$ which
202: characterizes the quantum
203: properties of radiation there is another parameter
204: \begin{equation}
205: \varrho=2\gamma^2\left\langle (\Delta\textbf{v} )^2\right\rangle,
206: \label{1.1}
207: \end{equation}
208: where $\left\langle (\Delta\textbf{v} )^2\right\rangle=
209: \left\langle \textbf{v}^2 \right\rangle -\left\langle \textbf{v} \right\rangle^2$
210: and $\left\langle\ldots \right\rangle$ denotes averaging over time.
211: In the case $\varrho \ll 1$ the radiation is of a dipole nature and it is formed
212: during the time of the order of the period of motion. In the case
213: $\varrho \gg 1$ the radiation is of magnetic bremsstrahlung nature and
214: it is emitted from a small part of the trajectory.
215:
216: In a crystal the parameter $\varrho$ depends on the angle incidence $\vartheta_0$
217: which is the angle between an axis (a plane) of crystal and the momentum
218: of a particle. If $\vartheta_0 \leq \vartheta_c$ (where $\vartheta_c \equiv
219: (2V_0/\varepsilon)^{1/2},~V_0$ is the scale of continuous potential of
220: an axis or a plane relative to which the angle $\vartheta_0$ is defined)
221: electrons falling on a crystal are captured into channels or low above-barrier
222: states, whereas for $\vartheta_0 \gg \vartheta_c$ the incident particles
223: move high above the barrier. In later case we can describe the motion using
224: the approximation of the rectilinear trajectory,
225: for which we find from Eq.(\ref{1.1}) the
226: following estimate
227: \begin{equation}
228: \varrho(\vartheta_0)=\left(2V_0/m\vartheta_0 \right)^2,
229: \label{2.1}
230: \end{equation}
231: For angles of incidence in the range $\vartheta_0 \leq \vartheta_c$
232: the transverse (relative to an axis or a plane) velocity of particle
233: is $v_{\perp}\sim \vartheta_c$ and the parameter obeys $\varrho \sim \varrho_c$
234: where
235: \begin{equation}
236: \varrho_c=2V_0\varepsilon/m^2.
237: \label{3.1}
238: \end{equation}
239: This means that side by side with the Lindhard angle $\vartheta_c$ the
240: problem under consideration has another characteristic angle
241: $\vartheta_v=V_0/m$ and $\varrho_c=(2\vartheta_v/\vartheta_c)^2$.
242:
243:
244: We consider here the photon emission in a thin crystal when the condition
245: $\varrho_c \gg 1$ is satisfied.
246: In this case the extremely difficult task of averaging of Eq.(\ref{1}) and
247: Eqs.(\ref{4})-(\ref{8}), derived for a given trajectory, over all
248: possible trajectories of electrons in a crystal simplifies radically.
249: In fact, if $\varrho_c \gg 1$ then in the range where trajectories are
250: essentially non-rectilinear ($\vartheta_0 \leq \vartheta_c,~v_{\perp} \sim
251: \vartheta_c$) the mechanism of photon emission is of the magnetic
252: bremsstrahlung nature and the characteristics of radiation can be expressed
253: in terms of local parameters of motion. Then the averaging procedure can be
254: carried out simply if one knows the distribution function in the transverse
255: phase space $dN(\varrho, \textbf{v}_{\perp})$, which for a thin crystal
256: is defined directly by the initial conditions of incidence of particle
257: on a crystal. For a given angle of incidence $\vartheta_0$ we have
258: $dN/N=d^3r F(\textbf{r}, \vartheta_0)/V$, where $V$ is the volume of a
259: crystal and $N$ is the total number of particles. In the axial case
260: the function $F(\textbf{r}, \vartheta_0)$ is of the form
261: \begin{equation}
262: F_{ax}(\mbox{\boldmath$\varrho$}, \vartheta_0)=\int
263: \frac{d^2 \varrho_0}{S(\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0))}
264: \vartheta((\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0)-
265: U(\mbox{\boldmath$\varrho$})),
266: \label{4.1}
267: \end{equation}
268: where $U(\mbox{\boldmath$\varrho$})$ is the continuous axial potential
269: dependent on the transverse coordinate $\mbox{\boldmath$\varrho$}$ normalized
270: so that $U(\mbox{\boldmath$\varrho$})=0$ at the boundary of a cell;
271: \begin{equation}
272: S(\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0))=
273: \int \vartheta((\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0)-
274: U(\mbox{\boldmath$\varrho$}))d^2\varrho,\quad
275: \varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0)=
276: \frac{\varepsilon \vartheta_0^2}{2} +
277: U(\mbox{\boldmath$\varrho$}_0),
278: \label{5.1}
279: \end{equation}
280: where $\vartheta(x)$ is the Heaviside function:$\vartheta(x)=0$ if $x < 0$
281: and $\vartheta(x)=1$ if $x > 0$.
282: For the planar case we have
283: \begin{equation}
284: F_{pl}(x, \vartheta_0)=\int \frac{\vartheta(\varepsilon_{\perp 0}-U(x))dx_0}
285: {v(\varepsilon_{\perp 0}, x)}
286: \left[\int \frac{\vartheta(\varepsilon_{\perp 0}-U(y))dy}
287: {v(\varepsilon_{\perp 0}, y)} \right]^{-1},
288: \label{6.1}
289: \end{equation}
290: where $U(x)$ is the continuous potential of plane dependent on the
291: coordinate $x$,~$v(\varepsilon_{\perp 0}, x)=
292: \left[2(v(\varepsilon_{\perp 0}-U(x))/\varepsilon \right]^{1/2}$
293: is the transverse velocity of an electron, and $\varepsilon_{\perp 0}
294: =\varepsilon \vartheta_0^2/2+U(x_0)$. It should be noted that in the
295: case when $\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0)>U_0$
296: (the above-barrier particles) the distribution Eq.(\ref{4.1}) becomes
297: uniform, whereas the distribution Eq.(\ref{6.1}) becomes
298: uniform only if $\varepsilon_{\perp 0} \gg U_0$, $U_0$ is the depth
299: of potential well.
300:
301: Substituting Eq.(\ref{8}) into Eq.(\ref{1}) we find after
302: integration by parts of terms ${\bf nv_{1,2}}~({\bf nv_{1,2}} \rightarrow 1)$
303: the general expression for energy loss of radiating electron
304: ($dE_{\xi} = \omega dw_{\xi}$)
305: \begin{eqnarray}
306: && dE_{\xi} = -\frac{\alpha m^2}{8 \pi^2}\frac{d^3k}{\varepsilon\varepsilon'}
307: \int \frac{d^3r}{V}F(\textbf{r}, \vartheta_0) \int e^{-iA}
308: \left[\varphi_1(\xi)+\frac{1}{4}\varphi_2(\xi)\gamma^2\left(\textbf{v}_1-
309: \textbf{v}_2\right)^2\right]dt_1dt_2,
310: \nonumber \\
311: && A=\frac{\omega \varepsilon}{2\varepsilon'}\int_{t_1}^{t_2}
312: \left[\frac{1}{\gamma^2}+(\textbf{n}-\textbf{v}(t))^2\right]dt,
313: \nonumber \\
314: &&\varphi_1(\xi)=1+\xi\frac{\omega}{\varepsilon}, \quad
315: \varphi_2(\xi)=(1+\xi)\frac{\varepsilon}{\varepsilon'}+
316: (1-\xi)\frac{\varepsilon'}{\varepsilon}.
317: \label{7.1}
318: \end{eqnarray}
319: where $\alpha=e^2=1/137$, the vector $\textbf{n}$ is defined
320: in Eq.(\ref{1}), the helicity of emitted photon $\xi$
321: is defined in Eq.(\ref{8}).
322:
323: The circular polarization of radiation is defined by the Stoke's parameter
324: $\xi^{(2)}$:
325: \begin{equation}
326: \xi^{(2)}=\Lambda (\mbox{\boldmath$\zeta$}{\bf v}),\quad
327: \Lambda = \frac{dE_+ - dE_-}{dE_+ + dE_-},
328: \label{7.1a}
329: \end{equation}
330: where the quantity $(\mbox{\boldmath$\zeta$}{\bf v})$ defines the longitudinal
331: polarization of the initial electrons, $dE_+$ and $dE_-$ is the energy
332: loss for $\xi$=+1 and $\xi$=-1 correspondingly. In the limiting case
333: $\omega \ll \varepsilon$ one has $\varphi_2(\xi) \simeq 2(1+\xi\omega/\varepsilon)
334: =2\varphi_1(\xi)$. So the expression for the energy loss $dE_{\xi}$ contains the
335: dependence on $\xi$ only as a common factor $\varphi_1(\xi)$ only. Substituting
336: in Eq.(\ref{7.1a}) one obtains the universal result independent of a particular
337: mechanism of radiation
338: \begin{equation}
339: \xi^{(2)}=\frac{\omega}{\varepsilon}(\mbox{\boldmath$\zeta$}{\bf v})
340: \label{7.2a}
341: \end{equation}
342:
343: The periodic crystal potential $U(\textbf{r})$ can be
344: presented as the Fourier series (see e.g.\cite{BKS}, Sec.8 )
345: \begin{equation}
346: U(\textbf{r})=\sum_{\textbf{q}}G(\textbf{q})e^{-i\textbf{q}\textbf{r}},
347: \label{8.1}
348: \end{equation}
349: where $\textbf{q}=2\pi(n_1, n_2, n_3)/l;~l$ is the lattice constant.
350: The particle velocity can be presented in a form
351: $\textbf{v}(t) =\textbf{v}_0+\Delta\textbf{v}(t)$, where
352: $\textbf{v}_0$ is the average velocity. If
353: $\vartheta_0 \gg \vartheta_c$, we find $\Delta\textbf{v}(t)$
354: using the rectilinear trajectory approximation
355: \begin{equation}
356: \Delta\textbf{v}(t)=-\frac{1}{\varepsilon}\sum
357: \frac{G(\textbf{q})}{q_{\parallel}}\textbf{q}_{\perp}
358: \exp [-i(q_{\parallel}t+\textbf{q}\textbf{r})],
359: \label{9.1}
360: \end{equation}
361: where $ q_{\parallel}=(\textbf{q}\textbf{v}_0),~
362: \textbf{q}_{\perp}=\textbf{q}-\textbf{v}_0(\textbf{q}\textbf{v}_0)$.
363: If $\varrho_c \gg 1$, there is a range of angles $\vartheta_0$ which
364: satisfies $\vartheta_c \ll \vartheta_0 \ll \vartheta_v$, where
365: both the magnetic bremsstrahlung description and the rectilinear
366: trajectory approximation are valid. However, since the magnetic
367: bremsstrahlung approach provides the same description for all
368: angles $\vartheta_0 \ll \vartheta_v$, the formula obtained in some
369: range of $\vartheta_0$ remain valid up to $\vartheta_0=0$. It is
370: known that in range $\vartheta_0 \geq \vartheta_v$
371: the rectilinear trajectory approximation is valid as well for
372: $\varrho_c \gg 1$. So, after integration in Eq.(\ref{7.1}) over
373: ${\bf u}={\bf n}-{\bf v}_0~(d^3k=\omega^2 d\omega d{\bf u})$ using
374: Eq.(\ref{9.1})
375: and passing to the variables $t, \tau:~t_1=t-\tau,~t_2=t+\tau$,
376: we obtain after simple calculations the general expression for
377: the intensity of radiation valid for any angle of incidence
378: $\vartheta_0$
379: \begin{eqnarray}
380: && dI_{\xi} \equiv \frac{dE_{\xi}}{dt}
381: = \frac{i\alpha m^2}{4 \pi\varepsilon^2}
382: \omega d\omega \int \frac{d^3r}{V}F(\textbf{r}, \vartheta_0)
383: \int \frac{d\tau}{\tau-i0}\Bigg[\varphi_1(\xi)-\varphi_2(\xi)
384: \nonumber \\
385: && \times \left(\sum_{\textbf{q}}
386: \frac{G(\textbf{q})}{m q_{\parallel}}\textbf{q}_{\perp}
387: \sin(q_{\parallel}\tau)e^{-i\textbf{q}\textbf{r}} \right)^2 \Bigg]
388: e^{-iA_2},
389: \label{10.1}
390: \end{eqnarray}
391: where
392: \begin{eqnarray}
393: && A_2=\frac{m^2\omega\tau}{\varepsilon\varepsilon'} \left[1+
394: \sum_{\textbf{q},\textbf{q}^{\prime} }\frac{G(\textbf{q})G(\textbf{q}')}
395: {m^2 q_{\parallel} q_{\parallel}'}
396: (\textbf{q}_{\perp}\textbf{q}_{\perp}')\Psi(q_{\parallel}, q_{\parallel}', \tau)
397: \exp [-i(\textbf{q}+\textbf{q}')\textbf{r}]\right]
398: \nonumber \\
399: && \Psi(q_{\parallel}, q_{\parallel}', \tau)=
400: \frac{\sin(q_{\parallel}+ q_{\parallel}')\tau}{(q_{\parallel}+
401: q_{\parallel}')\tau}-\frac{\sin(q_{\parallel}\tau)}{q_{\parallel}\tau}
402: \frac{\sin(q_{\parallel}'\tau)}{q_{\parallel}'\tau}.
403: \label{11.1}
404: \end{eqnarray}
405:
406:
407:
408:
409: \section{Radiation for $\vartheta_0 \ll V_0/m$
410: (magnetic \\ bremsstrahlung limit)}
411: \setcounter{equation}{0}
412:
413: The behavior of intensity $I$ Eq.(\ref{10.1}) for various entry angles and
414: energies is determined by the dependence on these parameters of the phase $A_2$
415: given Eq.(\ref{11.1}). Here we consider the axial case for
416: $\vartheta_0 \ll V_0/m \equiv \vartheta_v$. The direction of crystal axis we take as
417: $z-$axis of the coordinate system. The order of magnitude of the double sum in $A_2$
418: is $(G/m)^2(q_{\perp}/q_{\parallel})^2\Psi(q_{\parallel}, q_{\parallel}', \tau)$.
419: For the vector $\textbf{q}$ lying in the plane $(x,y)$ we introduce notation
420: $\textbf{q}_t$, for such vectors one has $q_z=0$ and the quantities in
421: Eq.(\ref{11.1}) can be estimated in the following way:
422: \begin{equation}
423: G(\textbf{q}) \sim V_0,\quad q_{\perp} \sim 1/a,\quad q_{\parallel} \sim \vartheta_0/a,
424: \label{1.2}
425: \end{equation}
426: where $a$ is the size of the region of action of the continuous potential. For all
427: remaining vectors $q_{\perp} \sim q_{\parallel} \sim 1/a$. Then the contribution
428: to the sum of the terms with $q_z \neq 0$ will be $\sim (V_0/m)^2\Psi \leq (V_0/m)^2$.
429: Since $(V_0/m)^2 \ll 1$ this contribution can be neglected. Thus, we keep in the sum
430: only terms with $\textbf{q}_t$ for which its value is $\sim (V_0/m\vartheta_0)^2\Psi$.
431: The large value of the phase $A_2$ leads to an exponential suppression of intensity
432: $I$. Therefore the characteristic value of the variable $\tau$ in the integral
433: Eq.(\ref{10.1}) (which have the meaning of the formation time (length) of the process)
434: will be adjusted in a such way that the large factor $(V_0/m\vartheta_0)^2$ will be
435: compensated by the function $\Psi(q_{\parallel}, q_{\parallel}', \tau)$, i.e. for
436: small entry angles the contribution gives region where $q_{\parallel}\tau \ll 1$.
437: Expanding the phase $A_2$ correspondingly we find an approximate expression for
438: $\vartheta_0 \ll \vartheta_v$
439: \begin{eqnarray}
440: && A_2 \simeq \frac{m^2\omega\tau}{\varepsilon\varepsilon'}\Bigg\{1-
441: \frac{\tau^2}{3} \sum_{\textbf{q}_t,\textbf{q}_t'}
442: G(\textbf{q}_t)G(\textbf{q}_t')\frac{(\textbf{q}_t\textbf{q}_t')}{m^2}
443: \exp\left[-i(\textbf{q}_t+\textbf{q}_t')\mbox{\boldmath$\varrho$} \right]
444: \nonumber \\
445: && \times \left[ 1-\frac{\tau^2}{10}\left((\textbf{q}_t\mbox{\boldmath$\nu$})^2
446: + (\textbf{q}_t'\mbox{\boldmath$\nu$})^2
447: +\frac{2}{3} (\textbf{q}_t\mbox{\boldmath$\nu$})
448: (\textbf{q}_t'\mbox{\boldmath$\nu$})\right) \right]\Bigg\},
449: \label{2.2}
450: \end{eqnarray}
451: here $\mbox{\boldmath$\nu$}=\textbf{v}_0/v_0$ is the direction of entry of the
452: initial electron, $\mbox{\boldmath$\varrho$}=\textbf{r}_t$. We cam rewrite
453: Eq.(\ref{2.2}) in the terms of the average potential of atomic string
454: $\displaystyle{U(\mbox{\boldmath$\varrho$})=\sum_{\textbf{q}_t} G(\textbf{q}_t)
455: \exp(-i\textbf{q}_t \mbox{\boldmath$\varrho$})}$:
456: \begin{equation}
457: A_2 = \frac{m^2\omega\tau}{\varepsilon\varepsilon'}
458: \left\lbrace 1+\frac{\tau^2}{3} \textbf{b}^2\tau^2 +
459: \frac{\tau^4}{15} \left[(\textbf{b}(\mbox{\boldmath$\nu$}
460: \mbox{\boldmath$\nabla$})^2 \textbf{b})
461: +\frac{1}{3} ((\mbox{\boldmath$\nu$}
462: \mbox{\boldmath$\nabla$})\textbf{b})^2 \right] \right\rbrace,
463: \label{3.2}
464: \end{equation}
465: where $\textbf{b}=\mbox{\boldmath$\nabla$}U(\mbox{\boldmath$\varrho$})/m,~
466: \mbox{\boldmath$\nabla$}=\partial/\partial \mbox{\boldmath$\varrho$}$.
467: For the pre-exponential factor in Eq.(\ref{10.1}) we find
468: \begin{equation}
469: [\ldots] \simeq \varphi_1(\xi)-\varphi_2(\xi)\tau^2\left[\textbf{b}^2 +
470: \frac{\tau^2}{3} (\textbf{b}(\mbox{\boldmath$\nu$}
471: \mbox{\boldmath$\nabla$}) \textbf{b}) \right]
472: \label{4.2}
473: \end{equation}
474: Taking the integral over $\tau$ we obtain the spectral intensity
475: for $\vartheta_0 \ll V_0/m$.
476: \begin{eqnarray}
477: && dI^F_{\xi}(\omega)=\frac{\alpha m^2\omega d\omega}{2\sqrt{3}\pi\varepsilon^2}
478: \int \frac{d^2\varrho}{s} \Bigg\{\int \frac{d^2\varrho_0}
479: {s(\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0))}
480: \vartheta((\varepsilon_{\perp}(\mbox{\boldmath$\varrho$}_0)-
481: U(\mbox{\boldmath$\varrho$})) R_0(\lambda)
482: \nonumber \\
483: && -\frac{(\textbf{b}(\mbox{\boldmath$\nu$}
484: \mbox{\boldmath$\nabla$})^2 \textbf{b})}{3\textbf{b}^4}
485: R_1(\lambda)+
486: \frac{\lambda}{30\textbf{b}^4}\left[((\mbox{\boldmath$\nu$}
487: \mbox{\boldmath$\nabla$})\textbf{b})^2 +3 (\textbf{b}(\mbox{\boldmath$\nu$}
488: \mbox{\boldmath$\nabla$})^2 \textbf{b})\right]R_2(\lambda)\Bigg\},
489: \label{5.2}
490: \end{eqnarray}
491: where
492: \begin{eqnarray}
493: \hspace{-15mm}&& R_0(\lambda)=\varphi_2(\xi)K_{2/3}(\lambda)-
494: \varphi_1(\xi)\int_{\lambda}^{\infty} K_{1/3}(y)dy,
495: \nonumber \\
496: &&R_1(\lambda)=
497: \varphi_2(\xi)\left(K_{2/3}(\lambda)-\frac{2}{3\lambda}K_{1/3}(\lambda)\right),
498: \nonumber \\
499: \hspace{-15mm}&& R_2(\lambda)=\varphi_1(\xi)\left( K_{1/3}(\lambda)
500: -\frac{4}{3\lambda}K_{2/3}(\lambda)\right)
501: \nonumber \\
502: &&+\varphi_2(\xi)\left(\frac{4}{\lambda}K_{2/3}(\lambda)
503: -\left(1+\frac{16}{9\lambda^2} \right)K_{1/3}(\lambda) \right),
504: \label{6.2}
505: \end{eqnarray}
506: here $\displaystyle{\lambda=\frac{2m^2\omega}{3\varepsilon\varepsilon'|\textbf{b}|}}$,
507: $K_{\nu}(\sigma)$ is the modified Bessel function (McDonald's function).
508: Since the expression for $dI^F$ is independent of $z$, it follows that
509: $\int d^3r/V \rightarrow \int d^2\varrho/s$, where $s$ is the transverse cross
510: section area per axis. The term in Eq.(\ref{5.2}) with $R_0(\lambda)$
511: represent the spectral intensity in the magnetic bremsstrahlung limit
512: with the flux redistribution taken into account. The other terms are the
513: correction proportional $\vartheta_0^2$.
514:
515: If the potential $U(\mbox{\boldmath$\varrho$})$ can be considered as
516: axially symmetric, we put $U=U(\mbox{\boldmath$\varrho$}^2)$ and one
517: can integrate over angles of vector $\mbox{\boldmath$\varrho$}$. We
518: obtain
519: \begin{eqnarray}
520: && dI_{\xi}^F(\omega)=\frac{\alpha m^2\omega d\omega}{2\sqrt{3}\pi\varepsilon^2}
521: \int_{0}^{x_0} \frac{dx}{x_0} \Bigg\{\int_{0}^{x_0} \frac{dy}
522: {y_0(\varepsilon_{\perp}(y))}
523: \vartheta((\varepsilon_{\perp}(y)-
524: U(x)) R_0(\lambda)
525: \nonumber \\
526: && -\frac{1}{6}\left( \frac{m \vartheta_0}{V_0}\right)^2\Bigg[
527: \frac{xg''+2g'}{xg^3} R_1(\lambda)
528: \nonumber \\
529: && -\frac{\lambda}{20g^4x^2}\left(2x^2g^{\prime 2}+g^2+14gg^{\prime}x
530: +6x^2gg^{\prime \prime}\right)R_2(\lambda)\Bigg] \Bigg\},
531: \label{7.2}
532: \end{eqnarray}
533: where we have adopted a new variable $x=\mbox{\boldmath$\varrho$}^2/a_s^2,
534: ~x \leq x_0,~x_0^{-1}=\pi a_s^2dn_a=\pi a_s^2/s,~a_s$
535: is the effective screening radius
536: of the potential of the string, $n_a$ is the density of atoms in a crystal,
537: $d$ is the average distance between atoms of a chain forming the axis;
538: \begin{equation}
539: \varepsilon_{\perp}(y)=\frac{\varepsilon\vartheta_0^2}{2}+U(y),\quad
540: y_0(\varepsilon_{\perp}(y))=\int_{0}^{\infty}
541: \vartheta(\varepsilon_{\perp}(y)-U(x))dx.
542: \label{8.2}
543: \end{equation}
544: The notation $U^{\prime}(x)=V_0g(x)$ is used in Eq.(\ref{7.2}) and
545: \begin{equation}
546: \lambda=\frac{m^3a_s\omega}{3\varepsilon\varepsilon'V_0g(x)\sqrt{x}}=
547: \frac{u}{3\chi_s g(x)\sqrt{x}},~\chi_s=\frac{V_0\varepsilon}{m^3 a_s},
548: ~u=\frac{\omega}{\varepsilon'}
549: \label{9.2}
550: \end{equation}
551: For specific calculation we use the following for the potential of axis:
552: \begin{equation}
553: U(x)=V_0\left[\ln\left(1+\frac{1}{x+\eta} \right)-
554: \ln\left(1+\frac{1}{x_0+\eta} \right) \right].
555: \label{10.2}
556: \end{equation}
557: For estimates one can put $V_0 \simeq Ze^2/d,~\eta \simeq 2u_1^2/a_s^2$,
558: where $Z$ is the charge of the nucleus,
559: $u_1$ is the amplitude of thermal vibrations,
560: but actually the parameter of potential were determined by means of a fitting
561: procedure using the potential Eq.(\ref{8.1}) (table of parameters for
562: different crystals is given in Sec.9 of \cite{BKS}).
563: For this potential
564: \begin{equation}
565: g(x)=\frac{1}{x+\eta} - \frac{1}{x+\eta +1} = \frac{1}{(x+\eta)(x+\eta +1)}
566: \label{11.2}
567: \end{equation}
568:
569: We shall assume, for the sake of simplicity, that the distribution is uniform
570: over the transverse coordinates. This is true in the case of large angles of
571: incidence $\vartheta_0 > \vartheta_c$ and is approximately correct in the case
572: of beams with a large angular spread $\Delta \vartheta_0 \sim \vartheta_c$.
573: In this case and for the assumptions
574: $u \ll \chi_s,~(\chi_s/u)^{4/3}/\varrho(\vartheta_0)\ll 1$
575: we have
576: \begin{equation}
577: \frac{dI_{\xi}^M}{\omega d\omega}=\frac{\alpha m^2}{6\sqrt{3}\pi x_0 \varepsilon^2}
578: \Phi_{\xi}
579: \label{12.2}
580: \end{equation}
581: where
582: \begin{equation}
583: \Phi_{\xi}=\Gamma\left(\frac{2}{3}\right)\left(\frac{6\chi_s}{u}\right)^{2/3}
584: \left[\varphi_2(u)\left(\ln\frac{18\sqrt{3}\chi_s}{u} - \frac{\pi}{2\sqrt{3}}-
585: C -l_1(\eta)\right)-\frac{3}{2}\varphi_1(u)\right]
586: \label{13.2}
587: \end{equation}
588: here
589: \begin{eqnarray}
590: &&\varphi_1(u)=1+\frac{\xi u}{1+u}, \quad \varphi_2(u)=(1+\xi)(1+u)+\frac{1-\xi}{1+u},
591: \quad C=0.577216...,
592: \nonumber \\
593: && l_1(\eta)=3.975 \beta^{2/3}\left(1+\frac{8}{15}\beta
594: +\frac{7}{18}\beta^2\right)-\beta\left(\frac{3}{2}+\frac{9}{8}\beta
595: +\frac{13}{14}\beta^2\right),
596: \label{14.2}
597: \end{eqnarray}
598: where $\beta=\eta/(1+\eta)$. In derivation of Eq.(\ref{12.2}) the integration
599: over $x$ is carried out in the interval $(0, \infty)$. Since the main (logarithmic)
600: contribution into the integral in Eq.(\ref{7.2}) comes from the interval $x \sim
601: (\chi_s/u)^{2/3}$, it is clear that the asymptotic expression Eq.(\ref{12.2})
602: is valid only is $(\chi_s/u)^{2/3} < x_0$.
603:
604: At high energies ($\chi_s \gg 1$) the expression for the intensity
605: of radiation in the magnetic bremsstrahlung limit
606: $I_{\xi}^M(\varepsilon)$ becomes
607: \begin{equation}
608: I_{\xi}^M(\varepsilon) \simeq \frac{\alpha V_0 g_1\varepsilon}{2m x_0 a_s \chi_s^{1/3}}
609: \left[\left(1+\frac{11}{16}\xi\right)(\ln \chi_s +g_0 )- \frac{231}{512}\xi\right],
610: \label{15.2}
611: \end{equation}
612: where
613: \begin{eqnarray}
614: && g_1=\left(\frac{2}{3}\right)^6 6^{2/3}\Gamma\left(\frac{2}{3}\right)\simeq 0.3925
615: \nonumber \\
616: && g_0=\frac{\pi}{2\sqrt{3}}+\frac{5}{2}\ln 3 +\ln 2 -C -l_1(\eta)
617: \simeq 0.6756-l_1(\eta).
618: \label{16.2}
619: \end{eqnarray}
620:
621: The circular polarization of radiation is defined by Eq.(\ref{7.1a}),
622: where we substitute $dE_{\pm} \rightarrow dI_{\pm}$. In the hard end of spectrum
623: $(\varepsilon' \ll \varepsilon)~\Lambda \simeq 1$
624: since $dI_+ \propto \varepsilon/\varepsilon'$ and
625: $dI_- \propto \varepsilon'/\varepsilon$.
626: In the limit $\chi_s \gg 1$ the radiation spectrum extends into hard part
627: up to $u=\omega/(\varepsilon- \omega) \sim \chi_s a_s/u_1$, where under condition
628: $|(\mbox{\boldmath$\zeta$}{\bf v})| \simeq 1$ the emitted photons have
629: practically the complete circular polarization (the helicity coincide with
630: electron helicity).
631: For the integral radiation intensity (where the main contribution gives
632: the region $\omega \sim \varepsilon-\omega$) the circular polarization
633: follows from Eq.(\ref{15.2}):
634: \begin{equation}
635: \Lambda = \frac{11}{16}\left(1- \frac{21}{32(\ln \chi_s + g_0)}\right).
636: \label{18.2}
637: \end{equation}
638: Here the main term coincide with circular polarization of radiation
639: of polarized electrons in external electromagnetic field, see Eq.(4.88)
640: in \cite{BKS}.
641:
642: In Fig.1 the spectral intensities of radiation
643: $dI^M_+/d\omega$,$~dI^M_-/d\omega$ and the
644: spectral intensity of radiation of unpolarized electrons
645: $dI^M/d\omega=dI^M_+/d\omega + dI^M_-/d\omega$ in tungsten crystal,
646: axis $<111>$,
647: temperature $T=293~K$ are plotted for energy $\varepsilon=$250~GeV
648: (curves 1,2,3, respectively) and for energy $\varepsilon=$1~TeV
649: (curves 4,5,6, respectively). It is seen that when $\omega/\varepsilon
650: \rightarrow 1$ the intensity $dI^M_+/d\omega$ dominates
651: ($dI^M_+/d\omega \gg dI^M_-/d\omega$), while for
652: $\omega/\varepsilon \ll 1$ the degree of polarization diminishes
653: ($dI^M_+/d\omega \rightarrow dI^M_-/d\omega$).
654: It should be noted that Eq.(\ref{7.2a}) is
655: fulfilled within accuracy better 10\% for $\omega/\varepsilon \leq 0.2$.
656: In Fig.2 the circular polarization $\xi^{(2)}$ of radiation is plotted
657: versus $\omega/\varepsilon$ for the same crystal. This curve is true
658: for both energies:$\varepsilon=$250~GeV and $\varepsilon=$1~TeV.
659: Actually this means that it is valid for any energy in high-energy region.
660: At $\omega/\varepsilon$=0.8 one has $\xi^{(2)}$=0.94 and at
661: $\omega/\varepsilon$=0.9 one has $\xi^{(2)}$=0.99.
662:
663: It is instructive to compare the obtained results with bremsstrahlung of
664: polarized electrons in amorphous medium which one write as (see e.g.\cite{BKF})
665: \begin{equation}
666: \varepsilon \frac{dI_{\xi}^{C}}{d\omega}=
667: \frac{\alpha m^2 \varepsilon}{8 \pi \varepsilon_e}\left[
668: (1+\xi)\left(1-\frac{2}{3}\frac{\varepsilon'}{\varepsilon}\right)
669: +\frac{\varepsilon'^2}{\varepsilon^2}\left(1-\frac{\xi}{3}\right)
670: +\frac{2}{9L_1}\frac{\varepsilon'}{\varepsilon}
671: \left(1+\xi\frac{\omega}{\varepsilon}\right)\right],
672: \label{19.2}
673: \end{equation}
674: where $\varepsilon_e=m(8\pi Z^2\alpha^2 n_a \lambda_c^3L_1)^{-1}$,
675: for notations see Eqs.(\ref{7.2}) and (\ref{10.2}), $\lambda_c=1/m$
676: is the electron Compton wavelength, $L_1=2(\ln 183Z^{1/3} - f(Z\alpha)),~
677: f(z)={\rm Re}[\psi(1+iz)-\psi(1)]$, $\psi(x)$ is the logarithmic
678: derivative of the gamma function, the function $f(Z\alpha)$ gives
679: the Coulomb corrections, for tungsten
680: $L_1$=6.99, $\varepsilon_e=2.73~$TeV. The dependence of the circular
681: polarization $\xi^{(2)}$ on $\omega/\varepsilon$ for the
682: intensity Eq.(\ref{19.2}) is very close to shown in Fig.2.
683: For the integral bremsstrahlung intensity Eq.(\ref{19.2})
684: one has $\Lambda=5/9$.
685: However, as one can be seen from Fig.1, in the region
686: $\omega \sim \varepsilon$ the magnitude of intensity
687: in oriented crystal is around 10 times larger than given by
688: Eq.(\ref{19.2}) for both energies:$\varepsilon=$250~GeV and
689: $\varepsilon=$1~TeV.
690: The value $r=(I^M/I^C)_{max}= (L_{rad}/L_{ch})_{max}$ for different
691: crystals is discussed in Sec.17.3 and is given in Table 17.1 in
692: \cite{BKS}. This value attains maximum value for light elements:
693: in diamond for axis $<111>$ $r$=168, in silicon for axis $<110>$ $r$=81.
694:
695: It should be mentioned that for energies $\varepsilon \geq \varepsilon_e$
696: the process of hard photon emission ($\omega \sim \varepsilon$) is
697: affected by the multiple scattering (the LPM effect). At
698: $\varepsilon \gg \varepsilon_e$ the suppression of spectral intensity
699: of bremsstrahlung become essential for the whole spectrum
700: \cite{BK3}, \cite{BK}.
701: On the contrary, the influence of multiple scattering on the photon
702: emission in a oriented crystal is rather weak because the formation length
703: in a oriented crystal is much shorter than in an amorphous medium
704: (see Sec. 21 in \cite{BKS}).
705:
706:
707:
708: \section{Modified theory of coherent bremsstrahlung}
709: \setcounter{equation}{0}
710:
711:
712: The estimates of double sum in the phase $A_2$ made at the beginning of previous
713: section: $\sim (\vartheta_v/\vartheta_0)^2\Psi$ remain valid also for
714: $\vartheta_0 \geq \vartheta_v$, except that now the factor in the double sum is
715: $(\vartheta_v/\vartheta_0)^2 \leq 1$, so that the values
716: $|q_{\parallel}\tau| \sim 1$ contribute. We consider first the limiting case
717: $\vartheta_0 \gg \vartheta_v$, then this factor is small and $\exp(-iA_2)$
718: can be expanded accordingly. As a result Eq.(\ref{10.1}) acquires the
719: form
720: \begin{eqnarray}
721: && dI_{\xi}^{coh}(\omega)=-\frac{i\alpha m^2\omega d\omega}{4\pi\varepsilon^2}
722: \int_{-\infty}^{\infty} \frac{d\tau}{\tau-i0}
723: \exp\left(-i\frac{m^2\omega\tau}{\varepsilon\varepsilon'} \right)
724: \sum_{\textbf{q},\textbf{q}^{\prime} }\frac{G(\textbf{q})G(\textbf{q}')}
725: {m^2 q_{\parallel} q_{\parallel}'}
726: (\textbf{q}_{\perp}\textbf{q}_{\perp}')
727: \nonumber \\
728: && \left[\varphi_2(\xi)\sin(q_{\parallel}\tau) \sin(q_{\parallel}'\tau)
729: +i\varphi_1(\xi)\frac{m^2\omega\tau}{\varepsilon\varepsilon'}
730: \Psi(q_{\parallel}, q_{\parallel}', \tau)\right] \int \frac{d^3r}{V} \exp [-i(\textbf{q}+\textbf{q}')\textbf{r}].
731: \label{1.3}
732: \end{eqnarray}
733: The integration over coordinate $\textbf{r}$ in Eq.(\ref{1.3}) is elementary and
734: gives $\delta_{\textbf{q}+\textbf{q}', 0}$, after which the sum over $\textbf{q}'$
735: and the integrals over $\tau$ are easily calculated by means of the theory of
736: residues. Finally we obtain
737: \begin{eqnarray}
738: && dI_{\xi}^{coh}(\omega)=\frac{\alpha \omega d\omega}{8\varepsilon^2}
739: \sum_{\textbf{q}} |G(\textbf{q})|^2 \frac{\textbf{q}_{\perp}^2}{q_{\parallel}^2}
740: \left[\varphi_2(\xi) -\varphi_1(\xi)\frac{2m^2\omega}
741: {\varepsilon\varepsilon'q_{\parallel}^2}\left(|q_{\parallel}|-
742: \frac{m^2\omega}{2\varepsilon\varepsilon'}\right) \right]
743: \nonumber \\
744: && \times \vartheta\left(|q_{\parallel}|-
745: \frac{m^2\omega}{2\varepsilon\varepsilon'}\right).
746: \label{2.3}
747: \end{eqnarray}
748: The spectral distribution Eq.(\ref{2.3}) coincide with the result of standard
749: theory of coherent bremsstrahlung (CBS), see e.g. \cite{TM}.
750:
751: In the case $\chi_s \gg 1$ ($\chi_s$ is defined in Eq.(\ref{9.2})), one can
752: obtain from general expression Eq.(\ref{10.1}) the expression for spectral
753: intensity, the region of applicability of which is broader than that of
754: standard CBS theory. For this purpose it is necessary to take into account that
755: the phase $A_2$ Eq.(\ref{11.1}) has for $q_{\parallel}+ q_{\parallel}' \neq 0$
756: terms of the order $(\vartheta_v/\vartheta_0)^3/\chi_s$ and
757: $(\vartheta_v/\vartheta_0)^4/\chi_s^2$ which can be small even for
758: $\vartheta_0 \leq\vartheta_v$ if $\chi_s \gg 1$. Therefore, assuming that these
759: contributions are small, we carry out the corresponding expansion of $\exp(-iA_2)$,
760: while the term with $q_{\parallel}+ q_{\parallel}' = 0$ in the double sum in
761: $A_2$ will be retained in the exponent. As a result we obtain an expression
762: which coincides in the form with Eq.(\ref{1.3}) where we must make the
763: substitution
764: \begin{equation}
765: \exp\left(-i\frac{m^2\omega\tau}{\varepsilon\varepsilon'} \right) \rightarrow
766: \exp\left(-i\frac{m_{\ast}^2\omega\tau}{\varepsilon\varepsilon'} \right), \quad
767: m_{\ast}^2=m^2\left(1+\frac{\varrho}{2} \right).
768: \label{3.3}
769: \end{equation}
770: Above the parameter $\varrho$ (Eq.(\ref{1.1})) has the form
771: \begin{equation}
772: \frac{\varrho}{2}=\frac{1}{m^2}\sum_{\textbf{q},\textbf{q}'}G(\textbf{q})
773: G(\textbf{q}')
774: \frac{\textbf{q}_{\perp}\textbf{q}_{\perp}'}{q_{\parallel} q_{\parallel}'}
775: \left[\delta_{q_{\parallel}+ q_{\parallel}',0} -\delta_{q_{\parallel},0}
776: \delta_{q_{\parallel}',0}\right] =\sum_{\textbf{q},q_{\parallel} \neq 0}
777: \frac{|G(\textbf{q})|^2\textbf{q}_{\perp}^2 }{m^2q_{\parallel}^2},
778: \label{4.3}
779: \end{equation}
780: and in the term $\displaystyle{\frac{\sin(q_{\parallel}+ q_{\parallel}')\tau}
781: {(q_{\parallel}+ q_{\parallel}')\tau}}$ it is necessary to assume that
782: $q_{\parallel}+ q_{\parallel}' \neq 0$. The remaining calculations
783: are carried out in the same way as in the transition from Eq.(\ref{1.3}) to
784: Eq.(\ref{2.3}). The final result can be presented in a form
785: \begin{eqnarray}
786: && \frac{dI_{\xi}^{mcoh}(\omega)}{d\omega}=\frac{\alpha u}{8\varepsilon}
787: \sum_{\textbf{q}} |G(\textbf{q})|^2 \frac{\textbf{q}_{\perp}^2}{q_{\parallel}^2}
788: \Bigg[1+\xi +\frac{1-\xi}{(1+u)^2}
789: \nonumber \\
790: && -\frac{8[1+u(1+\xi)]}{(2+\varrho)(1+u)^2}\frac{u}{u_0}\left(1-
791: \frac{u}{u_0}\right) \Bigg] \vartheta(u_0-u),
792: \label{5.3}
793: \end{eqnarray}
794: where
795: \begin{equation}
796: u_0=\frac{4\varepsilon q_{\parallel}}{m^2 (2+\varrho)}, \quad
797: u=\frac{\omega}{\varepsilon'}
798: \label{5.3a}
799: \end{equation}
800: Equation (\ref{5.3}) is not more complicated than Eq.(\ref{2.3})
801: but has a significantly broader range of applicability.
802:
803: The spectral intensities Eqs.(\ref{2.3}) and (\ref{5.3}) can be
804: much higher than the Bethe-Heitler bremsstrahlung intensity for
805: small angles of incidence $\vartheta_0$ with respect to selected axis.
806: For the case $\vartheta_0 \ll 1$ the quantity $q_{\parallel}$
807: can be represented as
808: \begin{equation}
809: q_{\parallel} \simeq \frac{2\pi}{d}n+\textbf{q}_{\perp}\textbf{v}_{0.\perp}
810: \label{5.3b}
811: \end{equation}
812: The main contribution to the sum in Eqs.(\ref{2.3}) and (\ref{5.3}) for
813: small $\vartheta_0$ is given $\textbf{q}$ with $n=0$, then
814: \begin{equation}
815: q_{\parallel} \simeq \left(\frac{2\pi}{f}k \cos \varphi +
816: \frac{2\pi}{h}l \sin \varphi\right) \vartheta_0,
817: \label{5.3c}
818: \end{equation}
819: where $f$ and $h$ are the characteristic periods of the potential in the
820: transverse plane.
821:
822: Let us consider the spectral intensity of radiation in the extreme limit,
823: when the parameter $\lambda \equiv 2\varepsilon|q_{\parallel}|_{min}/m^2
824: \sim \varepsilon \vartheta_0/m^2 a_s \gg 1$. In this case the maximum
825: of intensity of coherent bremsstrahlung is attained at such values of
826: $\vartheta_0$ where the standard theory of coherent bremsstrahlung
827: becomes invalid. Bearing in mind that if $\lambda \gg 1$ and
828: $\vartheta_0 \sim V_0/m$, then $\chi_s \sim \lambda \gg 1$,
829: we can conveniently use a modified theory of coherent bremsstrahlung.
830:
831: The direction of transverse components of particle's velocity in
832: Eq.(\ref{5.3b}) can be selected in a such way, that
833: the spectral intensity described by Eq.(\ref{5.3}) has a sharp maximum
834: near the end of spectrum at
835: $\omega = 2\varepsilon\lambda(2+2\lambda+\varrho)^{-1}$
836: with relatively small (in terms of $\lambda^{-1}$) width
837: $\Delta \omega \sim \varepsilon(1+\varrho/2)/\lambda=
838: m^2(1+\varrho/2)/2 |q_{\parallel}|_{min}$:
839: \begin{equation}
840: \left(\frac{dI_{\xi}}{d\omega}\right)_{max}=
841: \frac{\alpha \varrho |q_{\parallel}|_{min}}{4(2+\varrho)}
842: \left(1+\xi + \frac{1-\xi}{(1+u_m)^2}\right),\quad
843: u_m=\frac{2\lambda}{2+\varrho}.
844: \label{8.3}
845: \end{equation}
846: It is seen that in the maximum of spectral distribution the radiation
847: intensity with opposite helicity ($\xi=-1$) is suppressed as
848: $1/(1+u_m)^2$. At $u > u_m$ one have to take into account the next
849: harmonics of particle acceleration. In this region of spectrum
850: the suppression of radiation intensity with opposite helicity is more strong,
851: so the emitted photons have nearly complete circular polarization.
852:
853: Comparing Eq.(\ref{8.3}) with Eq.(\ref{19.2}) for bremsstrahlung
854: for $\varepsilon' \ll \varepsilon$ we find that for the same circular
855: polarization ($\xi^{(2)} \simeq (\mbox{\boldmath$\zeta$}{\bf v})$) in the
856: particular case $\varrho = 1$ the magnitude of spectral intensity in
857: Eq.(\ref{8.3}) is about $\chi_s(\varepsilon_e)$ times larger than
858: the spectral intensity in Eq.(\ref{19.2}). For tungsten W one has
859: $\chi_s(\varepsilon_e)=78$. Taking into account that $\Delta \omega
860: \sim \varepsilon/\chi_s(\varepsilon) =
861: \varepsilon_e/\chi_s(\varepsilon_e)$ we find that
862: \begin{equation}
863: \Delta \omega \left(\frac{dI_{\xi}}{d\omega}\right)_{max}
864: \sim \frac{\varepsilon_e}{L_{rad}},\quad
865: \frac{dN_{\gamma}}{dt}
866: \sim \frac{\varepsilon_e}{\varepsilon}\frac{1}{L_{rad}},
867: \label{9.3}
868: \end{equation}
869: where $N_{\gamma}$ is the number of emitted photons.
870: So, it is seen from above analysis that under these conditions the
871: considered mechanism of emission of photons with circular polarization
872: is especially effective because there is gain both in monochromacity
873: of radiation and total yield of polarized photons near
874: hard end of spectrum.
875:
876:
877:
878: \section{Conclusions}
879: \setcounter{equation}{0}
880:
881: It is shown above that at high energy the radiation from
882: longitudinally polarized electrons in oriented crystals is circularly
883: polarized and $\xi^{(2)} \rightarrow 1$ near the end of spectrum.
884: This is true in magnetic bremsstrahlung limit $\vartheta_0 \ll V_0/m$
885: as well as in coherent bremsstrahlung region $\vartheta_0 > V_0/m$.
886: This is particular case of helicity transfer.
887:
888: It should be noted that in crossing channel: production of
889: electron-positron pair with longitudinally polarized particles
890: by the circularly polarized photon in an oriented crystal the same phenomenon
891: of helicity transfer takes place in the case when the final particle
892: takes away nearly all energy of the photon. The corresponding formulas can be
893: obtained from given above using substitution rules.
894: For processes in external electromagnetic
895: field this item is discussed in Sec.3.3 of \cite{BKS}.
896:
897: So, the oriented crystal is a very effective device for helicity transfer
898: from an electron to photon and back from a photon to electron or positron.
899: Near the end of spectrum this is nearly 100\% effect.
900:
901: \vspace{0.5cm}
902:
903: {\bf Acknowledgments}
904:
905: The authors are indebted to the Russian Foundation for Basic
906: Research supported in part this research by Grant
907: 03-02-16154.
908:
909:
910:
911:
912:
913:
914:
915:
916:
917:
918: \newpage
919: \begin{thebibliography}{99}
920: \bibitem{BK0} V. N. Baier, and V. M. Katkov,
921: Phys. Lett.,A {\bf 25} (1967) 492.
922: \bibitem{BK1} V. N. Baier, and V. M. Katkov,
923: Sov.Phys.JETP {\bf 26} (1968) 854.
924: \bibitem{BK2} V. N. Baier, and V. M. Katkov,
925: Sov.Phys.JETP {\bf 28} (1969) 807.
926: \bibitem{BKS} V. N. Baier, V. M. Katkov and V. M. Strakhovenko,
927: {\em Electromagnetic Processes at High Energies in Oriented
928: Single Crystals}, World Scientific Publishing Co, Singapore, 1998.
929: \bibitem{BLP} V. B. Berestetskii, E. M. Lifshitz and L. P. Pitaevskii,
930: {\em Quantum Electrodynamics} Pergamon Press, Oxford, 1982.
931: \bibitem{BKF} V. N. Baier, V. M. Katkov and V. S. Fadin,
932: {\em Radiation from Relativistic Electrons} (in Russian) Atomizdat,
933: Moscow, 1973.
934: \bibitem{BKS1} V. N. Baier, V. M. Katkov and V. M. Strakhovenko,
935: Sov.Phys.JETP {\bf 63} (1986) 467.
936: \bibitem{BKS2} V. N. Baier, V. M. Katkov and V. M. Strakhovenko,
937: Sov.Phys.JETP {\bf 65} (1987) 686.
938: \bibitem{TM} M. L. Ter-Mikaelian, {\em High Energy Electromagnetic
939: Processes in Condensed Media}, John Wiley \& Sons, 1972.
940: \bibitem{BK3} V. N. Baier and V. M. Katkov,
941: Phys.Rev. D {\bf 62} (2000) 036008.
942: \bibitem{BK} V. N. Baier and V. M. Katkov,
943: {\it ``Concept of formation length in radiation theory''}
944: hep-ph/0309211, 2003; Preprint BINP 2003-58, Novosibirsk, 2003.
945:
946: \end{thebibliography}
947:
948: \newpage
949:
950: {\bf Figure captions}
951:
952: \vspace{8mm}
953: \begin{itemize}
954:
955: \item {\bf Fig.1} Spectral intensity of radiation
956: $\displaystyle{\varepsilon\frac{dI^M(\omega)}{d\omega}}$ in tungsten crystal,
957: axis $<111>$, $T=293~K$ in units $\alpha m^2$ vs $\omega/\varepsilon$. For the
958: initial electron energy $\varepsilon$=250~GeV curve 1 is $dI^M_-/d\omega$,
959: curve 2 is $dI^M_+/d\omega$, curve 3 is $dI^M_+/d\omega+dI^M_-/d\omega$;
960: and for the
961: initial electron energy $\varepsilon$=1~TeV curve 4 is $dI^M_-/d\omega$,
962: curve 5 is $dI^M_+/d\omega$, curve 6 is $dI^M_+/d\omega+dI^M_-/d\omega$.
963:
964:
965: \item {\bf Fig.2} The circular polarization $\xi^{(2)}$
966: of radiation (for ($\mbox{\boldmath$\zeta$}{\bf v}$)=1)
967: vs $\omega/\varepsilon$ for the tungsten crystal,
968: axis $<111>$, $T=293~K$. The curve is valid for both energies:
969: $\varepsilon$=250~GeV and $\varepsilon$=1~TeV.
970:
971:
972:
973:
974:
975: \end{itemize}
976:
977: \end{document}
978: