Dependence of effective spectrum width of synchrotron radiation on particle energy

TL;DR

This paper introduces the concept of effective spectral width in synchrotron radiation, providing numerical calculations and analyzing how it depends on particle energy and initial harmonic, enhancing the quantitative understanding of spectral properties.

Contribution

It presents a new concept of effective spectral width and offers detailed numerical calculations for polarization components in classical synchrotron radiation theory.

Findings

Effective spectral width is numerically calculated for polarization components.

The dependence of spectral width and initial harmonic on particle energy is established.

Effective spectral width does not exceed 100 harmonics.

Abstract

For an exact quantitative description of spectral properties in the theory of synchrotron radiation, the concept of effective spectral width is introduced. In the classical theory, numeric calculations of effective spectral width (using an effective width not exceeding 100 harmonics) for polarization components of synchrotron radiation are carried out. The dependence of the effective spectral width and initial harmonic on the energy of a radiating particle is established.

Tables

Table 1. Table 1: Bondary harmonics and respective energy of effective spectral width for linear polarization components of SR and unpolarized radiation

$\Lambda$	${\nu }_2^{\left(1\right)}$	${\gamma }_2$	${\nu }_2^{\left(2\right)}$	${\tilde{\gamma }}_2$	${\nu }_3^{\left(1\right)}$	${\gamma }_3$	${\nu }_3^{\left(2\right)}$	${\tilde{\gamma }}_3$	${\nu }_0^{\left(1\right)}$	${\gamma }_0$	${\nu }_0^{\left(2\right)}$	${\tilde{\gamma }}_0$
\m

Table 2. Table 2: Bondary harmonics and respective energy of effective spectral width for circle polarization components of SR

$\Lambda$	${\nu }_1^{\left(1\right)}$	${\gamma }_1$	${\nu }_1^{\left(2\right)}$	${\tilde{\gamma }}_1$	${\nu }_{-1}^{\left(1\right)}$	${\gamma }_{-1}$	${\nu }_{-1}^{\left(2\right)}$	${\tilde{\gamma }}_{-1}$
˘

Equations

$$W_s=W\sum _{\nu =1}^{\infty }{\int }_0^{\pi }{f}_s(\beta ;\nu ,\theta )\sin \theta d\theta .$$

$$W=\frac{2}{3}\frac{c{e}^2}{R^2}({\gamma }^2-1{)}^2=\frac{2e^4{H}^2({\gamma }^2-1)}{3m_0^2{c}^3},\gamma =\frac{1}{\sqrt{1-{\beta }^2}},$$

$$f_2(\beta ;\nu ,\theta )=\frac{3{\nu }^2}{2{\gamma }^4}{J}_{\nu }^{\prime 2}(x);f_3(\beta ;\nu ,\theta )=\frac{3{\nu }^2}{2{\gamma }^4}\frac{{\cos }^2\theta }{{\beta }^2{\sin }^2\theta }{J}_{\nu }^2(x);$$

$$f_{\pm 1}(\beta ;\nu ,\theta )=\frac{3{\nu }^2}{4{\gamma }^4}{\left[J_{\nu }^{\prime }(x)\pm \epsilon \frac{\cos \theta }{\beta \sin \theta }{J}_{\nu }(x)\right]}^2;x=\nu \beta \sin \theta ;\epsilon =-\frac{e}{|e|};$$

$$f_0(\beta ;\nu ,\theta )=f_2(\beta ;\nu ,\theta )+f_3(\beta ;\nu ,\theta )=f_1(\beta ;\nu ,\theta )+f_{-1}(\beta ;\nu ,\theta ).$$

$$W_s=W\left[{\Phi }_s^{(+)}(\beta )+{\Phi }_s^{(-)}(\beta )\right];{\Phi }_s^{(+)}(\beta )=\sum _{\nu =1}^{\infty }{F}_s^{(+)}(\beta ;\nu ),{\Phi }_s^{(-)}(\beta )=\sum _{\nu =1}^{\infty }{F}_s^{(-)}(\beta ;\nu );$$

$$F_s^{(+)}(\beta ;\nu )={\int }_0^{\pi /2}{f}_s(\beta ;\nu ,\theta )\sin \theta d\theta ,F_s^{(-)}(\beta ;\nu )={\int }_{\pi /2}^{\pi }{f}_s(\beta ;\nu ,\theta )\sin \theta d\theta ,$$

$$F_s^{(-)}(\beta ;\nu )=F_s^{(+)}(\beta ;\nu ),{\Phi }_s^{(-)}(\beta )={\Phi }_s^{(+)}(\beta ),s=0,2,3;$$

$$F_1^{(-)}(\beta ;\nu )=F_{-1}^{(+)}(\beta ;\nu ),{\Phi }_1^{(-)}(\beta )={\Phi }_{-1}^{(+)}(\beta ),$$

$$F_{-1}^{(-)}(\beta ;\nu )=F_1^{(+)}(\beta ;\nu ),{\Phi }_{-1}^{(-)}(\beta )={\Phi }_1^{(+)}(\beta ).$$

$$F_2^{(+)}(\beta ;\nu )=\frac{3\nu }{4\gamma {}^4{\beta }^3}\left[2{\beta }^2{J}_{2\nu }^{\prime 2}{\int }_0^{2\nu \beta }{J}_{2\nu }(x)dx-2\nu \beta {\int }_0^{2\nu \beta }\frac{J_{2\nu }(x)}{x}dx\right],$$

$$F_3^{(+)}(\beta ;\nu )=\frac{3\nu }{4\gamma {}^4{\beta }^3}\left[2\nu \beta {\int }_0^{2\nu \beta }\frac{J_{2\nu }(x)}{x}dx-{\int }_0^{2\nu \beta }{J}_{2\nu }(x)dx\right],$$

$$F_0^{(+)}(\beta ;\nu )=F_2^{(+)}(\beta ;\nu )+F_3^{(+)}(\beta ;\nu )=\frac{3\nu }{4\gamma {}^4{\beta }^3}\left[2{\beta }^2{J}_{2\nu }^{\prime 2}){\int }_0^{2\nu \beta }{J}_{2\nu }(x)dx\right],$$

$$F_{\pm 1}^{(+)}(\beta ;\nu )=\frac{1}{2}{F}_0^{(+)}(\beta ;\nu )\pm \frac{3\nu J{}_{\nu }^2(\nu \beta )}{4\gamma {}^4{\beta }^2}.$$

$${\Phi }_2^{(+)}(\beta )=\frac{6+{\beta }^2}{16},{\Phi }_3^{(+)}(\beta )=\frac{2-{\beta }^2}{16},{\Phi }_0^{(+)}(\beta )=\frac{1}{2},{\Phi }_{\pm 1}^{(+)}(\beta )=\frac{1}{4}\left[1\pm \frac{3}{4}{\chi }_1(\beta )\right].$$

$${\chi }_1(0)=1,{\chi }_1(1)=\frac{4}{\pi \sqrt{3}}.$$

$${\tilde{\Phi }}_s(\beta ;{\nu }^{(1)},{\nu }^{(2)})=\sum _{\nu ={\nu }^{(1)}}^{{\nu }^{(2)}}{F}_s^{(+)}(\beta ;\nu ),1⩽{\nu }^{(1)}⩽{\nu }^{(2)}⩽\infty .$$

$${\Phi }_s^{(+)}(\beta )={\tilde{\Phi }}_s(\beta ;{\nu }^{(1)}=1,{\nu }^{(2)}=\infty );{\Phi }_s^{(+)}(\beta )>{\tilde{\Phi }}_s(\beta ;{\nu }^{(1)}⩾1,{\nu }^{(2)}<\infty ).$$

$${\tilde{\Phi }}_s(\beta ;{\nu }^{(1)},{\nu }^{(2)})⩾\frac{1}{2}{\Phi }_s^{(+)}(\beta ).$$

$${\tilde{\Phi }}_s(\beta ;{\nu }_s^{(1)}(\beta ),{\nu }_s^{(2)}(\beta ))-\frac{1}{2}{\Phi }_s^{(+)}(\beta )⩾0.$$

$${\Lambda }_s(\beta )={\nu }_s^{(2)}(\beta )-{\nu }_s^{(1)}(\beta )+1,{\Lambda }_s(\beta )⩾1.$$

$$P_s(\beta ;\nu )=\frac{F_s^{(+)}(\beta ;\nu )}{{\Phi }_s^{(+)}(\beta )}.$$

$$\sum _{\nu =1}^{\infty }{P}_s(\beta ;\nu )=1.$$

$$\sum _{\nu ={\nu }_s^{(1)}(\beta )}^{{\nu }_s^{(2)}(\beta )}{P}_s(\beta ;\nu )-\frac{1}{2}⩾0.$$

$$\sum _{l=0}^{{\Lambda }_s-1}{P}_s(\beta ;{\nu }_s^{(1)}+l)-\frac{1}{2}=0.$$

$$P_2(\beta ;k)-\frac{1}{2}=0.$$

$$1⩽\gamma ⩽{\gamma }_2(1,1),$$

$$P_2(\beta ;k)+P_2(\beta ;k+1)-\frac{1}{2}=0.$$

$$\gamma >{\gamma }_2(1,1),$$

$${\gamma }_2(1,1)<\gamma ⩽{\gamma }_2(2,1),$$

Full text

Dependence of effective spectrum width of synchrotron radiation on particle

energy

V. G. Bagrov , D. M. Gitman , A. D. Levin , A. S. Loginov , A. D. Saprykin Tomsk State University, Russia; Institute of High Current Electronics, SB RAS, Tomsk, Russia; Institute of Physics, University of Sao Paulo, Brazil. E-mail: bagrov@phys.tsu.ruTomsk State University, Russia; Institute of Physics, University of Sao Paulo, Brazil; P.N. Lebedev Physical Institute, Moscow, Russia. E-mail: gitman@if.usp.br Institute of Physics, University of Sao Paulo, Brazil. E-mail:alevin@if.usp.brTomsk State University, Russia. E-mail: as_loginov@phys.tsu.ru Tomsk State University, Russia. E-mail: alexx.saprykin.phys@gmail.com

Abstract

For an exact quantitative description of spectral properties in the theory of synchrotron radiation, the concept of effective spectral width is introduced. In the classical theory, numeric calculations of effective spectral width (using an effective width not exceeding 100 harmonics) for polarization components of synchrotron radiation are carried out. The dependence of the effective spectral width and initial harmonic on the energy of a radiating particle is established.

1 Introduction

As one of the major quantitative characteristics of spectral distributions for electromagnetic radiation, one commonly uses the concept of spectral half-width. For spectral distributions having a sharp maximum, spectral half-width is the most informative physical characteristic.

However, once a spectral distribution has no pronounced maximum, spectral half-width ceases to be an adequate quantitative characteristic. In particular, this is exactly the case of spectral distributions for synchrotron radiation (SR), and therefore SR spectral half-width has neither been calculated theoretically, nor measured experimentally.

Instead of spectral half-width, the present study proposes to introduce a new precise quantitative characteristic of SR spectral distributions: effective spectral width. It is shown how this quantity can be calculated theoretically, and which physically relevant information can be obtained using this quantity.

At present, the theory of synchrotron radiation is a fairly well-developed area of theoretical physics. Its main elements are described in monographs (e.g. [1, 2, 3, 4, 5, 6, 7]) and numerous articles. Among the most important SR physical features, one should take into account a high polarization degree of radiation and a unique structure of spectral distributions. All theoretically predictable SR properties have been confirmed by experiment. The development of SR theory allows one not only to predict radiation characteristics qualitatively, but also to offer exact quantitative characteristics of physically important properties. For example, the high degree of SR polarization was qualitatively predicted by theory more than half a century ago (see, e.g.,[1]); precise quantitative characteristics for linear polarization were also obtained, whereas such quantitative characteristics for circular polarization were obtained much later [8, 9, 10, 11].

In order to set up the problem, we now present some well-known expressions of the classical SR theory for the physical characteristics of synchrotron radiation, which can be found in [1, 2, 3, 4, 5, 6, 7].

The spectral-angular distribution for radiation power of SR polarization components can be written as

[TABLE]

Here, the following notation is used: $\theta$ is the angle between the control magnetic field strength and the radiation field pulse; $\nu$ is the number of an emitted harmonic; the charge orbital motion rate is $v=c\beta$, where $c$ is the speed of light; $W$ is the total radiated power of unpolarized radiation, which can be written as

[TABLE]

where $e$ is the particle charge; $R$ is the orbit radius; $H$ is the control field strength; $m_{0}$ is the charge rest mass; $\gamma$ is the relativistic factor. The index $s$ numbers the polarization components: $s=2$ corresponds to the $\sigma$-component of linear polarization; $s=3$ corresponds to the $\pi$-component of linear polarization; $s=1$ corresponds to right-hand circular polarization; $s=-1$ corresponds to left-hand circular polarization; $s=0$ corresponds to the power of unpolarized radiation. The functions $f_{s}(\beta;\,\nu,\,\theta)$ have the form [1, 2, 3, 4, 5, 6, 7]

[TABLE]

[TABLE]

[TABLE]

Here, $J_{\nu}(x)$ and $J_{\nu}^{\prime}(x)$ are the Bessel functions and their derivatives. In what follows, the case of an electron is considered, which corresponds to $\varepsilon=1$.

2 Spectral distribution for polarization components of synchrotron

radiation in the upper half-space

It is well known [1, 2, 3, 4, 5, 6, 7] that the angle range $0\leqslant\theta<\pi/2$ (this range will be called the upper half-space) is dominated by right-hand circular polarization, and the angle range $\pi/2<\theta\leqslant\pi$ (this range will be called the lower half-space) is dominated by left-hand circular polarization (exact quantitative characteristics of SR properties were first obtained in [8, 9, 10, 11]). However, if we integrate in (1. 1) over $\theta\ (0\leqslant\theta\leqslant\pi)$, then the differences in the spectral distribution of right- and left-hand circular polarizations disappear. To reveal these differences, the expressions (1. 1) can be represented as

[TABLE]

[TABLE]

and it suffices to study the properties of functions $F_{s}^{(+)}(\beta;\,\nu)$ (respectively, the properties of functions $\Phi_{s}^{(+)}(\beta)$), due to the evident relations

[TABLE]

[TABLE]

[TABLE]

Integration over $\theta$ in the upper half-space $0\leqslant\theta\leqslant\pi/2$ in (2. 1) can be carried out exactly, yielding the expressions

[TABLE]

[TABLE]

[TABLE]

[TABLE]

The sums over the harmonics $\nu$ in (2. 1) can also be calculated exactly, yielding the expressions

[TABLE]

The function $\chi_{1}(\beta)$ introduced above was defined and studied in [8]. In particular, it was shown [8] that in the segment $0\leqslant\beta\leqslant 1$ the function $\chi_{1}(\beta)$ is finite and decreases monotonously; at the end of this segment, it takes the following values:

[TABLE]

3 Effective spectral width for polarization components of

synchrotron radiation

As one of the quantitative characteristics of physical properties for spectral distributions of SR polarization components, it is proposed to introduce the concept of effective spectral width $\Lambda_{s}(\beta)$. Let us define $\Lambda_{s}(\beta)$ as follows.

For each fixed value of $\beta$, we examine the quantities

[TABLE]

The following relations are obvious:

[TABLE]

Consider the set of values $\nu^{(1)},\,\,\nu^{(2)}\,(1\leqslant\nu^{(1)}\leqslant\nu^{(2)}<\infty)$ such that satisfy the inequality

[TABLE]

Obviously, such values $\nu^{(1)},\,\nu^{(2)}$ do exist for any $\beta$ (for example, $\nu^{(1)}=1$ necessarily yields such a finite value $\nu^{(2)}$). It is equally obvious that condition (3. 3) alone is generally insufficient to determine the pair of values $\nu^{(1)},\,\nu^{(2)}$. Let us now choose such $\nu_{s}^{(1)}(\beta)$, $\,\nu_{s}^{(2)}(\beta)$ that the condition (3. 3) should provide the minimum of the difference $\nu_{s}^{(2)}(\beta)-\nu_{s}^{(1)}(\beta)$, as well as the minimum of the non-negative value

[TABLE]

The effective spectral width $\Lambda_{s}(\beta)$ is defined by the expression

[TABLE]

Consequently, effective spectral width is the minimum spectral range that accounts for at least half of all the radiated power of a given polarization component. The harmonics $\nu_{s}^{(1)}(\beta)$ and $\nu_{s}^{(2)}(\beta)$ determine the beginning and the end of this minimum spectral range.

A definition equivalent to the one presented above for effective spectral width can be given using the concept of partial contributions $P_{s}(\beta;\nu)$ for individual spectral harmonics, introduced in [12]. Namely, we suppose, according to [12],****

[TABLE]

Then (2. 1) implies the property

[TABLE]

We choose some values $\nu_{s}^{(1)}(\beta)$ and $\nu_{s}^{(2)}(\beta)$ such that the minimum difference $\nu_{s}^{(2)}(\beta)-\nu_{s}^{(1)}(\beta)$ should provide the minimum of the non-negative value

[TABLE]

Introducing $\Lambda_{s}(\beta)$ in accordance with (3. 5), we arrive at the following equivalent definition: effective spectral width is the minimum spectral range at which the sum of partial contributions for individual harmonics is not less than $1/2$.

In practice, the most interesting case is the ultra-relativistic limit ($\beta\approx 1$, equivalent to $\gamma\gg 1$). In this case, the analytical study of effective spectral width and other physically interesting quantitative characteristics for spectral distributions of SR polarization components can be significantly extended. This study was carried out in [13].

Given a particular value of $\beta$ (or $\gamma$), it is a purely computational task to obtain the exact values of $\Lambda_{s}(\beta)$ and $\nu_{s}^{(1)}(\beta)$. In this article, we present a numerical study of the region $1\leqslant\Lambda_{s}(\beta)\leqslant 100$. It is essential to observe the following. The effective width $\Lambda_{s}(\beta)$ is a positive integer, so there exists a range of $\beta$ (corresponding to a range of $\gamma$; hereinafter, we only indicate $\gamma$) in which $\Lambda_{s}(\beta)$ is constant.

4 Analysis of numerical results for effective spectral width of

synchrotron radiation

The main results of a numerical study for effective spectral width of SR polarization components are given by our Table.

The numerical study is carried out as follows. For each type of polarization $s$, we examine the sequences of integers $\Lambda_{s}=1\,,2\,,3\,...$ and $\nu_{s}^{(1)}=1\,,2\,,3\,...$ (it is evident that $\nu_{s}^{(2)}=\nu_{s}^{(1)}+\Lambda_{s}-1$) and determine the regions of values $\gamma_{s}$ for which the condition (3. 8) (equivalent to (3. 4)) is satisfied. It is clear that the boundary points of possible regions for $\gamma_{s}$ can be found, according to (3. 8), as solutions of the equations

[TABLE]

The roots $\beta_{s}=\beta_{s}(\Lambda_{s},\nu_{s}^{(1)})$ of these equations determine the boundary points $\gamma_{s}=\gamma_{s}(\Lambda_{s},\nu_{s}^{(1)})$, in accordance with (1. 2).

Let us consider in more detail the calculation method and the results pertaining to the $\sigma$- component of linear polarization. These results are given in the column $s=2$ of our Table. We denote $\nu_{2}^{(1)}=k=1\,,2\,,3\,...$.

At the first step, we examine the smallest possible value $\Lambda_{2}=1$. In this case, equations (4. 1) have the form

[TABLE]

According to the results of [12], all partial contributions at $k\geqslant 2$ are such that $P_{2}(\beta;k)<1/2$ for all the values of $\gamma$, whereas in the first harmonic ($k=1$) such a region of values $\gamma$ does exist, and equation (4. 2) at $k=1$ has a single root, $\gamma_{2}=\gamma_{2}(1\,,1)$, shown in the Table.

Consequently, in the region of values $\gamma$

[TABLE]

for the $\sigma$-component of linear polarization, the effective spectral width is $\Lambda=1$, where $\nu_{2}^{(1)}=1$. There are no other values of $\tilde{\nu}_{2}^{(1)}$ and $\tilde{\gamma}_{2}$ for $\Lambda=1$, which is also indicated in the Table.

Next, we examine the value $\Lambda_{2}=2$. In this case, equations (4. 1) have the form

[TABLE]

Since at the previous step it has been established that in the region $1\leqslant\gamma\leqslant\gamma_{2}(1\,,1)$ the effective width is $\Lambda_{2}=1$, it is required to examine the solutions of equations (4. 4) only in the region

[TABLE]

Analysis of equations (4. 4) shows that these equations have a unique solution, $\gamma_{2}(2\,,1)$, only for $k=1$, where $\gamma_{2}(1\,,1)<\gamma_{2}(2\,,1)$.

Consequently, in the region of values $\gamma$

[TABLE]

the effective width is $\Lambda_{2}=2$, where $\nu_{2}^{(1)}=1$. There are no other values of $\tilde{\nu}_{2}^{(1)}$ and $\tilde{\gamma}_{2}$ for $\Lambda=2$.

At the next step, we examine the value $\Lambda_{2}=3$. Equations (4. 1) at $\Lambda_{2}=3$ have the form

[TABLE]

and we should only be concerned with solutions of these equations that belong to the region

[TABLE]

since at the previous steps it has been found that the effective spectral width at $1\leqslant\gamma\leqslant\gamma_{2}(2\,,1)$ is such that $\Lambda_{2}\leqslant 2$. It turns out that under restriction (4. 8) equations (4. 7) possess the solutions $\gamma_{2}(3\,,1)$ at $k=1$, and $\gamma_{2}(3\,,2)$ at $k=2$. These solutions obey the inequalities $\gamma_{2}(2\,,1)<\gamma_{2}(3\,,2)<\gamma_{2}(3\,,1)$.

Consequently, in the region of values $\gamma$

[TABLE]

the effective width is $\Lambda_{2}=3$, where $\nu_{2}^{(1)}=1$. However, inside the region (4. 9) there is a subregion:

[TABLE]

which admits a larger value, $\nu_{2}^{(1)}=\tilde{\nu}_{2}=2$ , as shown in the Table.

At the next step, we examine $\Lambda_{2}=4\,,5\,,6$. For these values of $\Lambda_{2}$ the results of calculation are similar to those obtained for $\Lambda_{2}=3$, namely, these are the values of $\gamma$ that belong to the region

[TABLE]

with the effective width being equal to the corresponding $\Lambda_{2}$, where $\nu_{2}^{(1)}=1$. However, inside the region (4. 11) there is a subregion:

[TABLE]

which admits a larger value, $\nu_{2}^{(1)}=\tilde{\nu}_{2}=2$, as shown in the Table.

At the next steps, we examine $\Lambda_{2}=7\div 11$, and each step to follow involves solutions of equations (4. 1) only in the region of values $\gamma$ being larger than those at the previous step. For these values $\Lambda_{2}$, equations (4. 1) possess the solutions $\gamma_{2}(\Lambda_{2},\nu_{2}^{(1)})$ corresponding to the three values $\nu_{2}^{(1)}=1$, $\nu_{2}^{(1)}=2$, $\nu_{2}^{(1)}=\tilde{\nu}_{2}=3$, and satisfying the inequalities

[TABLE]

In the Table, we indicate the smallest value $\nu_{2}^{(1)}=1$ and the respective largest value $\gamma_{2}(\Lambda_{2},1)$, as well as the largest value $\nu_{2}^{(1)}=\tilde{\nu}_{2}=3$ and the respective smallest value $\gamma_{2}(\Lambda_{2},3)$. For $\Lambda_{2}=7\div 11$, the corresponding regions of $\gamma$ are determined by relation (4. 11), but now there are two subregions: one is determined by relation (4. 12), so that the values of $\gamma$ admit $\nu_{2}^{(1)}=2$; the other one is determined by the condition

[TABLE]

and this region of $\gamma$ also admits the value $\nu_{2}^{(1)}=3$.

The next step concerns the region $\Lambda_{2}=12\div 27$. Here, it is essential that the smallest possible value is $\nu_{2}^{(1)}=2$; otherwise, the results are identical to the case $\Lambda_{2}=7\div 11$.

In general, the column $s=2$ of the Table indicates for each $\Lambda_{2}$ the smallest possible value $\nu_{2}^{(1)}=\nu_{2}^{(1)}(\Lambda_{2})$ and the corresponding largest value $\gamma_{2}=\gamma_{2}(\Lambda_{2},\nu_{2}^{(1)}(\Lambda_{2}))$, as well as the largest possible value $\tilde{\nu}_{2}$ for this $\Lambda_{2}$ and the corresponding smallest value $\tilde{\gamma}_{2}$. Possible intermediate values $\nu_{2}^{(1)}$ between the smallest $\nu_{2}^{(1)}(\Lambda_{2})$ and the largest $\tilde{\nu}_{2}$, as well as the respective intermediate values $\gamma_{2}$, are not specified. The intermediate values of $\gamma_{2}$ always satisfy relations (4. 13).

Consequently, in the column $s=2$ we indicate the regions of values $\gamma$

[TABLE]

for which the effective spectral width for the $\sigma$-component of SR linear polarization equals to $\Lambda_{2}$. We also indicate the initial points of the effective spectral width. These points are not determined uniquely. For the smallest initial value $\nu_{2}^{(1)}$, the range of values (4. 15) taken by $\gamma$ is the largest one, while this region is the smallest one for the largest possible value $\tilde{\nu}_{2}.$

For the other polarization components, the results of calculation are given in the respective columns of the Table. In particular, the Table shows that for equal values $\Lambda_{s}$ the corresponding values $\gamma_{s}$ obey the inequalities

[TABLE]

At a fixed energy $\gamma$, the corresponding values of $\Lambda_{s}$ are restricted by

[TABLE]

In this way, for each polarization component of synchrotron radiation we have found energy regions at which the effective spectral width equals to $\Lambda_{s}$, and the initial harmonic of this effective width is determined. Numeric calculations have been carried out in the case $\Lambda_{s}\leqslant 100$.

In the ultrarelativistic case, the corresponding results have been obtained in [13].

\DTLsettabseparator\DTLloaddb

[noheader,keys = a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,r,q,s,t,u,omitlines = 0]datatable_out.dat

\DTLdeletedb

data

Acknowledgements

The work of Bagrov and Gitman is partially supported by RFBR research project No. 15- 02-00293a and by Tomsk State University Competitiveness Improvement Program. Gitman thanks CNPq and FAPESP for their permanent support. Levin thanks CNPq for permanent support.