- About this Journal ·
- Abstracting and Indexing ·
- Advance Access ·
- Aims and Scope ·
- Article Processing Charges ·
- Articles in Press ·
- Author Guidelines ·
- Bibliographic Information ·
- Citations to this Journal ·
- Contact Information ·
- Editorial Board ·
- Editorial Workflow ·
- Free eTOC Alerts ·
- Publication Ethics ·
- Reviewers Acknowledgment ·
- Submit a Manuscript ·
- Subscription Information ·
- Table of Contents

Advances in Numerical Analysis

Volume 2014 (2014), Article ID 231498, 11 pages

http://dx.doi.org/10.1155/2014/231498

## Nonlinear Double-Layer Bragg Waveguide: Analytical and Numerical Approaches to Investigate Waveguiding Problem

Department of Mathematics and Supercomputing, Penza State University, Krasnaya Street 40, Penza 440026, Russia

Received 30 June 2013; Accepted 14 November 2013; Published 22 January 2014

Academic Editor: Theodore E. Simos

Copyright © 2014 Yury G. Smirnov et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### Abstract

The paper is concerned with propagation of surface TE waves in a circular nonhomogeneous two-layered dielectric waveguide filled with nonlinear medium. The problem is reduced to the analysis of a
nonlinear integral equation with a kernel in the form of the Green function. The existence of propagating TE waves for chosen nonlinearity (the Kerr law) is proved using the contraction mapping method. Conditions under which *k* waves can propagate are obtained, and intervals of localization of the corresponding propagation constants are found. For numerical solution of the problem, a method based on solving an auxiliary Cauchy problem (the shooting method) is proposed. In numerical experiment, two types of nonlinearities are considered and compared: the Kerr nonlinearity and nonlinearity with saturation. New propagation regime is found.

#### 1. Introduction

Theory of circle cylindrical dielectric waveguides attracts attention for a long time. Linear theory of such waveguides is known for years; see, for example, [1–4]. At the same time it is well known that the permittivity of a dielectric, in general, depends nonlinearly on the intensity of an electromagnetic field; see [5, 6]. For this reason, the linear theory can be applied only for fields of low intensity. However, for applications, sometimes it is necessary to raise the intensity of the field, for example, in order to compensate the losses. What happens in the case of high intensity (when the permittivity of the dielectric depends nonlinearly on the intensity of the field)? Is it possible to preserve waveguide regimes and, if so, how to determine propagation constants in the nonlinear case? It is not always easy to answer these questions. However, in the case of “simple” geometry (circle cylindrical and plane-layered waveguides) and polarized (TE and TM) electromagnetic waves, it is possible to answer these questions.

To the best of our knowledge, the first rigorous study of polarized electromagnetic wave propagation in a nonlinear circle cylindrical dielectric homogeneous waveguide is in [7, 8]. Then, there were several works, where some important cases for nonlinear but homogeneous permittivity have been investigated; see [9–12]. The next step was to apply earlier developed technique to the cases of multilayered waveguides and inhomogeneous nonlinear permittivity.

In the paper [13], we considered integral equation approach to derive dispersion equations in a nonlinear waveguiding problem. This approach can be used for numerical implementation. However, there exists a faster and simpler numerical approach that allows determining propagation constants in a wide range of waveguiding problems. Here, we are going to demonstrate this numerical approach for special waveguiding problem. As Bragg (or multilayered) waveguides have application (see, e.g., [14, 15]), we demonstrate here that the approach from [13] can be easily extended to be applied to more complicated problems. In order to justify the analytical approach given here, we will widely use the results of the paper [13].

#### 2. Statement of the Problem

Let us consider a three-dimensional space with cylindrical coordinate system . The space is filled with an isotropic medium of dielectric permittivity , where is the permittivity of free space, without sources. The medium is assumed to be isotropic and nonmagnetic. A cylindrical dielectric waveguide of circular cross section is presented as

and is placed in with its generating line parallel to the axis . It is supposed that everywhere , where is the permeability of free space.

The cross section of the waveguide, which is perpendicular to its axis, consists of two concentric circles of radii and ; that is, the waveguide is two-layered (see Figure 1): is the radius of the inner cylinder and is the thickness of the outer cylindrical shell. In the case of linear homogeneous media, such multilayered waveguides were studied in [3]. A practical application of a nonlinear Bragg waveguide is pointed out in [15].

The geometry of the problem is shown in Figure 1. The waveguide is unlimitedly continued in direction.

Let and be complex amplitudes of an electromagnetic field. The complex amplitudes , must satisfy Maxwell’s equations [7] have continuous tangential field components on the media interface , and obey the radiation condition at infinity; that is, the electromagnetic field decays exponentially as in the region ; is the circular frequency.

The permittivity in the entire space has the form , where

and ; is the orthonormal vector in the direction; is the Euclidean scalar product; ; is a real constant; , are real positive constants such that

#### 3. TE Waves

Let us consider TE-polarized waves in the monochromatic mode (see [16]): where are complex amplitudes and where is unknown spectral parameter.

Let . Thus, substituting components (6) into (2) and using the notation , we obtain where and is defined by formula (3).

We assume that function is sufficiently smooth:

We will seek under condition .

Let and . In the domains and , equation (7) takes the forms, respectively,

In the domain , equation (7) takes the form where , , .

The necessary solutions to (9) and (10) must be written in the following form:

The functions and are the modified Bessel functions and and are constants. The radiation condition is fulfilled because exponentially as (see [17]).

#### 4. Transmission Conditions and Transmission Problem

Transmission conditions for the functions and result from the continuity conditions for the tangential field components and have the form where is the jump of the limit values of the function at a point .

Let us formulate the transmission eigenvalue problem (problem ) to which the problem of surface waves propagating in a cylindrical waveguide has been reduced. The goal is to find quantities such that, for given (or ), there is a nonzero function that is defined by formulas (12) and (13) for and , respectively, and solves equation (11) for ; moreover, the function thus defined satisfies conditions (8) and transmission conditions (14).

The quantities solving problem are called *eigenvalues*, and the corresponding functions are called *eigenfunctions*. Such definition of an eigenvalue was given in [18] and found application later in similar nonlinear problems (see, e.g., [19, 20]). It should be noted that the eigenvalue depends on the value of the eigenfunction on one of the waveguide boundaries.

#### 5. Nonlinear Integral Equation

Let us consider the equation (linear part of (11)) where , .

Let . Consider the boundary value problem

Let be a complete system of (real) eigenvalues and orthonormalized with weight eigenfunctions of boundary value problem (16) (see [21]); this system exists as coefficients of the equation satisfy all necessary conditions. Then, for , the boundary value problem has only the trivial solution. This means that, for , there uniquely exists the Green function of the boundary value problem

Near an eigenvalue , the Green function can be presented in the form (see, e.g., [22, 23]) where the series converges absolutely and uniformly for all and for all from any compact set, which does not contain , . The function is regular in the neighborhood of , while are the above-mentioned systems of eigenvalues and normalized in eigenfunctions, see [24]. For big numbers it is true that the eigenvalue (see [22]).

Using the same technique as in [13], we obtain an integral representation for the solution of (11) for : where .

Assume that and . From the properties of the Green function, it follows that the kernel is a continuous function in the square , .

It follows from the results of the paper [13] that the nonlinear operator where , is completely continuous on each bounded subset of .

Consider the cubic equation , where

It can be shown (see [16]) that the following assertions hold, which will be needed below.

Proposition 1. *Under the condition
**
The equation has two nonnegative solutions
**
such that .*

Proposition 2. *Under the condition , (20) has at least one solution such that .*

Using Propositions 1 and 2, it can be shown (see [16]) that the following theorems hold.

Theorem 3. *If , where
**
then (20) has a unique continuous solution: such that .*

Theorem 4. *Let the kernel and the right-hand side f of integral equation (20) be continuous functions of the parameter , , , on a certain interval of the real axis. Let also
**Then, the solution to (20), for , exists, is unique, and continuously depends on the parameter , .*

We should note that the statement of the theorem holds for sufficiently small .

Approximate solutions of integral equation (20), written in the form , can be found using the iterative process , , with . The iterative procedure together with the convergence theorem is given in [13]. These iterated solutions can be applied to determine approximated eigenvalues of problem .

It should be noted that there is a simpler and more efficient approach to determine approximate eigenvalues and eigenfunctions which is based on the Cauchy problem method [20, 25] (shooting method) with initial data taken, for example, at the point . Of course, when performing the calculations, we need to specify the constant ; in fact, this means specifying the eigenfunction at the point . As was noticed previously, the eigenvalues in the nonlinear problem depend on the initial conditions. The iterative algorithm described above yields an approximate solution but is too cumbersome for implementation. The main purpose of the theoretical apparatus used here is to rigorously prove the existence of eigenvalues. The numerical results presented below were obtained by the Cauchy problem method.

#### 6. Dispersion Equation

Setting and in (20), using transmission conditions (14), and taking into account solutions (12) and (13), we obtain

Expressing the constant from the first and second equations of the system and equating the resulting expressions, we obtain a dispersion equation of the form where

#### 7. Existence of Solutions of the Dispersion Equation

The zeros of the function are the values of for which problem has a nontrivial solution; that is, if is such that , then the eigenvalues of the problem are determined by the equation .

Let us analyze the existence of solutions of the dispersion equation for the linear problem. From (28), with , we obtain . This equation can be rewritten as where and we took into account that .

Using (19) one finds As series in (19) converges absolutely and uniformly it is possible to reorder terms of the series. So, after some algebra one can find that

It is supposed that the eigenvalues are ordered in the way that and so, it is true that if .

Using derived formula (30) can be rewritten in the following form where , , and The left-hand side of (35) is bounded and continuous in the vicinity of the pole .

Let us consider the behavior of the right-hand side of (35) in the vicinity of the same pole.

As one can easily see that

It follows from this consideration that at least one root of the equation lies between adjacent poles (or, what the same is, adjacent eigenvalues of problem (16)) and .

The fact that the term in the expression , where , does not vanish can be easily proved using the theorem of the uniqueness of a solution to the Cauchy problem [13].

Now, we can show that the solutions of the equation exist.

Indeed, let there be integers and , such that where ; , are eigenvalues of problem (16). Above we rearranged eigenvalues in the order of magnitude.

Choose sufficiently small numbers such that the Green function exists and is continuous on the union of the intervals

In addition, we assume that are chosen such that the inequalities are valid.

In other words, we chose in such a way which move away from the poles of the Green function but preserve solutions of the linear problem (for ).

Hence, is bounded. Moreover, the product can be made sufficiently small by choosing a suitable . Consider the dispersion equation . The function is continuous and changes its sign when varies from to . Since is bounded for varying from to , by choosing a suitable , we can always make the equation have at least roots , such that , , where are chosen on .

The main result of the present work is the following theorem.

Theorem 5. *Let the numbers , , and satisfy the condition and let there be integers and such that
**
where are the eigenvalues of boundary value problem (16). Then, there is a number such that, for any , problem has at least eigenvalues such that , .*

*Proof. *The Green function exists for all . Clearly,
is a continuous function of . Let and let . According to Theorems 3, and 4, there is a unique solution to equation (13) for any . This solution is a continuous function, and . Let . Estimating , we obtain , where is a constant.

The function is continuous, and the equation has at least one root inside the interval , . Let , . Then, is positive and independent of .

If , then

The function is continuous; hence, the equation has a root inside ; that is, . We can choose .

Theorem 5 implies that, under the above-formulated conditions, there exist symmetric TE-polarized waves propagating without attenuation in nonhomogeneous cylindrical dielectric circular waveguides filled with a nonmagnetic isotropic Kerr nonlinear medium. This result generalizes the well-known assertion for both homogeneous dielectric waveguides (see [3]) and nonhomogeneous dielectric circular waveguides (see [4]) filled with a linear medium (with ).

#### 8. Existence of Propagation Constants

Consider the Cauchy problem for with the initial conditions where the values and are determined from (12) and have the following forms: where is supposed to be known.

It is clear from the foregoing that the calculation necessitates specified constant or . The constant can be specified on any of the waveguide boundaries. Note that, in contrast to problems for linear media, the propagation constants (eigenvalues) in problems for nonlinear media depend on the constant at one of the boundaries.

We use the classic results of the theory of ordinary differential equations on the existence and uniqueness of the solution to the Cauchy problem and on the continuous dependence of the solution to the Cauchy problem on the parameter [24].

Let and let be a certain constant. We determine the set and number is such that , where is the vector of right-hand side of (46) if it is rewritten in normal form.

The following propositions are valid (see [26] for the proof).

Proposition 6. *Solution to Cauchy problem (46)-(47) exists for all , where , is unique, continuously depends on for all , and is continuously differentiable with respect to .*

*Remark 7. *Since we consider the case , it is evident that Proposition 6 has a nonlocal character; that is, we can always choose and such that solution exists and is continuous for and for all .

Let, for , the equality be fulfilled. If the equality is also fulfilled, then is an eigenvalue of problem .

Inherently, formulae (50) and (51) are transmission conditions (14) at the point .

In order to determine , suppose that (50) is fulfilled and consider the function

From transmission conditions (14) on the boundary and solution (13), we obtain

Using this system, we can rewrite in the following form:

It follows from this formula that the value is expressed in terms of values of the solution to the Cauchy problem only.

One can show [26, 27] that the following statement is valid.

Proposition 8. *Assume that the conditions of Proposition 6 are fulfilled and the segment is such that . Then, there is, at least, one propagation constant (one eigenvalue) of problem .*

*Remark 9. *The condition is sufficient for the existence of propagation constant of problem .

It follows from general results of the theory of ordinary differential equations that calculated eigenvalues of the problem are stable with respect to initial values.

#### 9. Determination of Approximate Propagation Constants

The method under consideration makes it possible to plot the dependence of (normalized) propagation constant on the (normalized) radius of the waveguide. Curves (or ), where is the circular frequency, are dispersion curves in such problems. If curve depends on the incident field amplitude (a circumstance we deal with in the considered problem), these curves are referred to as energy dispersion curves [6]. Since normalized variables are employed, the plot of the dependence is called a dispersion curve (or an energy dispersion curve).

Consider Cauchy problem (46)-(47). Let , be certain numbers. Assume that,

We split the segments and into and sections, respectively. Then, we have , and . Here, , , , and . Then, for each pair of indices , we have pair of initial values , where and . In spite of the fact that the initial values are independent of , the double indexing is suitable for our analysis.

Now, we can formulate the Cauchy problem for (46) with the initial condition . The solution of this problem yields the values

Suppose that and construct the function

Assume that, for given , there exist and such that

Hence, there exists that is an eigenvalue of the considered problem of wave propagation, and this eigenvalue is associated with the layer’s thickness . Value can be found within an arbitrary accuracy by means of, for example, the dichotomy method.

On the basis of the dichotomy method, we develop a technique for the determination of an approximate value of the propagation constant.

Let us specify error for the determination of the value of propagation constant . Let the segment be such that

The sought eigenvalue is and the approximate eigenvalue is .

Let us determine the center of the segment and calculate . We check the following conditions.(1)If , then is the sought approximate eigenvalue.(2)If , then . In this case, we set and ; hence, .(3)If , then . In this case, we set and , hence; .

We repeat dichotomic division times to obtain the approximate value . Evidently, . Let us choose number such that . Then, for example, the center of the segment , that is, , can be considered as approximate value of where .

One can show [26] that the following statement is valid.

Proposition 10. *Assume that the conditions of Proposition 8 are fulfilled and that, by means of the dichotomy method, sequence of approximate values of propagation constant is obtained; then, , where is an exact eigenvalue.*

#### 10. Numerical Results

Numerical results are found with the help of the method suggested in the previous section. Numerical experiments were carried out for two types of nonlinearities: the Kerr law and nonlinearity with saturation. For the inhomogeneity of the waveguide, the following functions, that specify the permittivities and in the layer , are used.(1)For ,(a);(b) (homogeneous case);(c).(2)For ,(a);(b) (homogeneous case);(c),

where is a positive real constant.

In figures below eigenvalues or dependence (see Figures 2–4 and 7–9) and eigenfunctions (see Figures 5, 6, 10, and 11) are shown.

In fact, Cases 1(b) and 2(b) are homogeneous; however, we consider them as inhomogeneous with constant inhomogeneity (it prevents us from mentioning each time the difference).

The following values of parameters are used for calculations given in Figures 2–11: , , , , and .

For the Kerr nonlinearity (Figures 2–6), .

For the nonlinearity with saturation (Figures 7–11), , and .

Dispersion curves for Cases 1(a), 1(b), and 1(c) are shown in Figures 2–4, respectively; dispersion curves for Cases 2(a), 2(b), and 2(c) are shown in Figures 7–9, respectively.

In Figures 2–4 and Figures 7–9, the vertical axis corresponds to and the horizontal axis corresponds to the thickness of the outer cylindrical shell (see Figure 1).

In Figures 5, 6, 10, and 11, the vertical axis corresponds to the value of the plotted function and the horizontal axis corresponds to the value .

Red lines in Figures 2–4 and 7–9 correspond to the linear inhomogeneous case and blue curves correspond to the nonlinear inhomogeneous case (in the same figure, the inhomogeneity is the same for both types of curves; red and blue curves differ from each other only by the nonlinear term).

Vertical dashed line in Figures 2, 3, 7, and 8 corresponds to ; in Figures 4 and 9 vertical dashed line corresponds to . Eigenvalues of the problem are points of intersections of this dashed line with the dispersion curves; first three points in Figures 2–4 are marked (the smallest value, red dot, corresponds to the linear case and the others two (green and black dots) correspond to the Kerr case); first four points in Figures 7–9 are marked (the smallest value, red dot, corresponds to the linear case and the other three (green, blue, and black dots) correspond to the case with saturated nonlinearity).

In Figures 5 and 6, eigenfunctions for Cases 1(a) and 1(c), respectively, are plotted. The color of a curve corresponds to the color of marked eigenvalue in Figures 2 and 4. The larger an eigenvalue is, the higher maximum corresponding eigenfunction has.

In Figures 10 and 11, eigenfunctions for the Cases 2(a) and 2(c), respectively, are plotted. The color of a curve corresponds to the color of marked eigenvalue in Figures 7 and 9. The larger an eigenvalue is, the higher maximum corresponding eigenfunction has.

If we compare red (linear case) and blue (nonlinear case) dispersion curves in Figures 2–4, we can notice that in the linear cases there is only finite number of the eigenvalues (this fact is well known) and for the Kerr case is we can suppose infinite number of the eigenvalues (as it is for the case of plane waveguide with, the Kerr nonlinearity; see, for example, [16, 28]). In each nonlinear case, there are eigenvalues that are close to corresponding linear case (these eigenvalues can be determined with the help of perturbation theory; existence of these very eigenvalues is proved analytically in this paper). At the same time, it is easy to see, from Figures 2–4 and 7–9, that in each nonlinear case there are eigenvalues that cannot be determined from perturbation theory (we call them “purely nonlinear” eigenvalues).

Calculation time usually is about 2–5 hours with a modern laptop (it depends on the grid density).

#### 11. Conclusion and Discussion

As it is obvious, the analytical method allows proving existence of eigenvalues in the nonlinear problem which are close to eigenvalues in the corresponding linear problem. However, as it is seen from the plots above, in the nonlinear problem, there are new eigenvalues that cannot be determined as a perturbation of eigenvalues in the linear problem. These new eigenvalues correspond to the new (“purely” nonlinear) propagation regime. At the same time, proposed numerical method allows finding all eigenvalues. Of course this method is efficient in the case of discrete eigenvalues.

In Figures 2–4, groups of three eigenvalues are marked. The red dot (smallest value) is an eigenvalue of the linear problem (with ). The green dot is an eigenvalue of the nonlinear problem. When one passes to the limit , then the green dot tends to the red dot. The black dot does not tend to an eigenvalue of the linear problem, when ; this eigenvalue is “purely” nonlinear and corresponds to a new type of nonlinear waves.

The same conclusion can be made for the eigenvalues shown in Figures 7–9. In these figures, groups of four eigenvalues are marked. Green, blue, and black dots correspond to the nonlinear problem. The green dot is a perturbation of the solution of the linear problem (red dot) and tends to the red dot, when . Blue and black dots are “purely” nonlinear eigenvalues; they also correspond to a new type of nonlinear waves.

Whether these mathematically predicted “purely” nonlinear waves really exist, it is a hypothesis that can be proved or disproved in the physical experiment.

The numerical method considered in the paper has the following advantages. (+) The method is easy to be implemented, because the Cauchy problem for a system of ordinary differential equations can be solved by means of the standard tools of any mathematical software package. (+) The method works substantially (“substantially” here means that the implementation of the numerical method based on auxiliary Cauchy problem is simpler and this method calculates faster approximately by a factor of ten.) faster than the numerical method based on the use of integral dispersion equations. (+) The method allows determining of approximate eigenvalues within an arbitrary given accuracy.

And the following is its disadvantage. (−) When is one of the eigenvalues of the problem, then the total derivative of function will respect to must not vanish at .

This method is also applied for plane-multilayered structures; see the paper “*electromagnetic wave propagation in nonlinear-layered waveguide structures: computational approach to determine propagation constants*” of Valovik in [29].

#### Conflict of Interests

The authors declarethat there is no conflict of interests regarding the publication of this paper.

#### Acknowledgments

The authors thank the referees of the paper for their valuable critique. The first author (Yu.S.) is supported by the Russian Foundation for Basic Research under project no. 12-007-97010-r-a; the second author (E.S.) is supported by the Russian Foundation for Basic Research under project no. 14-01-31234-mol_a; the third author (D. V.) is supported by the Russian Federation President Grant under project no. MK-90.2014.1.

#### References

- M. J. Adams,
*An Introduction to Optical Waveguide*, John Wiley & Sons, New York, NY, USA, 1981. - A. Snyder and J. Love,
*Optical Waveguide Theory*, Chapman & Hall, London, UK, 1983. - G. I. Veselov and S. B. Raevskii,
*Layered Metall-Dielectric Waveguides*, Radio i Svyaz, Moscow, Russia, 1988, (Russian). - A. S. Zil'bergleit and Yu. I. Kopilevich,
*Spektralnaya Teoriya Regulyarnykh Volnovodov*, FTI, Leningrad, Russia, 1983, (Russian). View at MathSciNet - L. D. Landau, E. M. Lifshitz, and L. P. Pitaevskii,
*Course of Theoretical Physics*, vol. 8 of*Electrodynamics of Continuous Media*, Butterworth-Heinemann, Oxford, UK, 1993. - N. N. Akhmediev and A. Ankevich,
*Solitons, Nonlinear Pulses and Beams*, Chapman & Hall, London, UK, 1997. - P. N. Eleonskii, L. G. Oganes’yants, and V. P. Silin, “Cylindrical nonlinear waveguides,”
*Journal of Experimental and Theoretical Physics*, vol. 35, no. 1, pp. 44–47, 1972. - P. N. Eleonskii and V. P. Silin, “Propagation of electromagnetic waves in an inhomogeneous nonlinear medium,”
*Journal of Experimental and Theoretical Physics*, vol. 39, no. 1, pp. 67–70, 1974. - Y. Smirnov, H. W. Schürmann, and Y. Shestopalov, “Integral equation approach for the propagation of TE-waves in a nonlinear dielectric cylindrical waveguide,”
*Journal of Nonlinear Mathematical Physics*, vol. 11, no. 2, pp. 256–268, 2004. View at Publisher · View at Google Scholar · View at Zentralblatt MATH · View at MathSciNet - Yu. G. Smirnov and S. N. Kupriyanova, “Propagation of electro-magnetic waves in cylindrical dielectric waveguides filled with a nonlinear medium,”
*Computational Mathematics and Mathematical Physics*, vol. 44, no. 10, pp. 1850–1860, 2004. View at MathSciNet - H. W. Schürmann, G. Yu. Smirnov, and V. Yu. Shestopalov, “Propagation of te-waves in cylindrical nonlinear dielectric waveguides,”
*Physical Review E*, vol. 71, no. 1, Article ID 016614, 10 pages, 2005. - Yu. G. Smirnov and E. A. Horosheva, “On the solvability of the nonlinear boundary eigenvalue problem for tm waves propagation in a circle cylyindrical nonlinear waveguide,”
*Izvestiya Vysshikh Uchebnykh Zavedenij. Povolzh. Region. Fiziki-Matematicheskie Nauki*, no. 3, pp. 55–70, 2010, (Russian). - Yu. G. Smirnov and D. V. Valovik, “On the problem of electromagnetic waves propagating along a nonlinear inhomogeneous cylindrical waveguide,”
*ISRN Mathematical Physics*, vol. 2013, Article ID 184325, 7 pages, 2013. View at Publisher · View at Google Scholar - J. D. Chatterton and J. L. Shohet, “Guided modes and loss in a plasma-filled Bragg waveguide,”
*Journal of Applied Physics*, vol. 102, no. 6, Article ID 063304, 2007. View at Publisher · View at Google Scholar · View at Scopus - Yu. Rapoport, A. Boardman, V. Grimalsky, et al., “Metamaterials for space physics and the new method for modeling isotropic and hyperbolic nonlinear concentrators,” in
*Proceedings of the 14th International Conference on Mathematical Methods in Electromagnetic Theory*, pp. 76–79, Ukraine, Kharkiv, August 2012. - Yu. G. Smirnov and D. V. Valovik,
*Electromagnetic Wave Propagation in Non-Linear Layered Waveguide Structures*, Penza State University Press, Penza, Russia, 2011. - A. F. Nikiforov and V. B. Uvarov,
*Special Functions of Mathematical Physics: A Unified Introduction with Applications*, Birkhäuser, Basel, Switzerland, 1988. View at MathSciNet - D. V. Valovik and Yu. G. Smirnov, “Propagation of TM waves in a Kerr nonlinear layer,”
*Computational Mathematics and Mathematical Physics*, vol. 48, no. 12, pp. 2217–2225, 2008. View at Publisher · View at Google Scholar · View at Scopus - D. V. Valovik, “Propagation of TM waves in a layer with arbitrary nonlinearity,”
*Computational Mathematics and Mathematical Physics*, vol. 51, no. 9, pp. 1622–1632, 2011. View at Publisher · View at Google Scholar · View at Scopus - D. V. Valovik and E. V. Zarembo, “The method of cauchy problem for solving a nonlinear eigenvalue transmission problem for tm waves propagating in a layer with arbitrary nonlinearity,”
*Computational Mathematics and Mathematical Physics*, vol. 53, no. 1, pp. 78–92, 2013. - R. Courant and D. Hilbert,
*Methods of Mathematical Physics*, vol. 1, Interscience Publishers, New York, NY, USA, 1953. - M. A. Naimark,
*Linear Differential Operators. Part I: Elementary Theory of Linear Differential Operators. Part II: Linear Differential Operators in Hilbert Space*, Frederick Ungar Publishing, New York, NY, USA, 1967. View at MathSciNet - V. S. Vladimirov,
*Equations of Mathematical Physics*, Mir, Moscow, Russia, 1984. - N. P. Erugin,
*Book for Reading on General Course of Differential Equations*, Nauka i Tekhnika, Minsk, Russia, 1979, Russian. View at MathSciNet - D. V. Valovik and E. V. Zarembo, “Solution of the nonlinear eigenvalue boundary-value problem for tm electromagnetic waves propagating in a kerrnonlinear layer by means of the cauchy problem method,”
*Journal of Communications Technology and Electronics*, vol. 58, no. 1, pp. 62–65, 2013. - E. Yu. Smol’kin, “Method of cauchy problem to solve a nonlinear transmission eigenvalue problem for tm wave propagating in a circle double-layer dielectric waveguide with kerr nonlinearity,”
*Izvestiya Vysshikh Uchebnykh Zavedenij. Povolzh. Region. Fiziki-Matematicheskie Nauki*, no. 4, pp. 49–58, 2012 (Russian). - D. V. Valovik and E. Yu. Smol’kin, “Numerical solution to the problem of electromagnetis tm wave propagation in a circle dielectric waveguide filled with nonlinear medium,”
*Izvestiya Vysshikh Uchebnykh Zavedenij. Povolzh. Region. Fiziki-Matematicheskie Nauki*, no. 3, pp. 29–37, 2012 (Russian). - Yu. G. Smirnov and D. V. Valovik, “Nonlinear effects of electromagnetic tm wave propagation in anisotropic layer with kerr nonlinearity,”
*Advances in Mathematical Physics*, vol. 2012, Article ID 609765, 21 pages, 2012. View at Publisher · View at Google Scholar - D. V. Valovik,
*Inverse Problems and Large Scale Computations*, vol. 52 of*Springer Proceeding in Mathematics & Statistics*, Springer, New York, NY, USA, 2013.