-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.tex
661 lines (523 loc) · 44.2 KB
/
main.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
\documentclass[aps,pre,showpacs,twocolumn]{revtex4}
\usepackage{amssymb}
\usepackage{graphicx}
\usepackage{epstopdf}
\usepackage{colordvi}
\usepackage{color}
\usepackage{dcolumn,amsmath,amsthm,amscd,amsfonts,amssymb,epsfig,graphics,graphicx,eucal}
\usepackage[dvipsnames]{xcolor}
\usepackage{comment}
\usepackage[section]{placeins}
\pdfoutput=1
\newcommand{\PT}{{\cal PT}}
\newcommand{\sn}{\mbox{sn}}
\newcommand{\dn}{\mbox{dn}}
\newcommand{\tomega}{\tilde{\omega}}
\DeclareMathOperator{\sech}{sech}
\begin{document}
\title{Localized modes in saturable Kerr media embedded in non-$\PT$-symmetric complex localized potentials}
\author{F. C. Moreira$^{1}$, S. B. Cavalcanti$^{1}$ }
\affiliation{
$^1$Universidade Federal de Alagoas, Campus A. C. Sim\~oes - Av. Lourival Melo Mota, s/n, Cidade Universit\'aria, Macei\'o - AL 57072-900, Brazil}
\date{\today}
\begin{abstract}
One-dimensional nonlinear wave propagation in a medium whose refractive index is represented by a local non-${\cal PT}$ complex potential, is theoretically investigated. We report on the existence of families of stable asymmetric spatial solitons in a saturable nonlinear medium in the presence of a gain/loss asymmetrical profile. A tunable parameter defines the deviation from the ${\cal PT}- $symmetric case, and a thorough study on the existence and stability of soliton families is carried out. The properties of the nonlinear modes bifurcating from the eigenvalue of the underlying linear problem are investigated. We obtain the remarkable result that even within an asymmetrical profile one still may obtain families of stable solitons below and above the phase transition where a collision between two real eigenvalues of the linear equation bifurcates into a pair of complex eigenvalues. The eigenvalue ranges in the power-eigenvalue diagrams for different gain/loss profiles are inspected and, it is found that the saturable nonlinearity severely restricts these ranges as the eigenvalues tend quite fast to an asymptotic profile as power increases. Examples of the dynamics of the asymmetric solitons obtained by numerical simulations are shown to agree quite well with the analytical results.
\end{abstract}
\pacs{42.65.Tg, 42.65.Sf }
\maketitle
%\input{epsf.tex}
%\epsfverbosetrue
\section{Introduction}
The past decade has witnessed an intense activity in linear optical systems, described by non-Hermitian Hamiltonians, triggered
by reports on a wide class of complex potentials that obeys the parity-temporal symmetry, the well known $\PT$ symmetry~\cite{Bender1}.
In spite of being non-Hermitian, these $\PT$-symmetric Hamiltonians exhibit real spectra, and also are probability conservative systems. These
Hamiltonians fit perfectly $\PT$-symmetric optical structures through the association of the local potential function with complex-valued refractive index profiles \cite{Ruter2010,peschel,peng}. This association
has been quite fruitful and has turned the field of optics to be a particularly adequate ground to test the amazing consequences of $\PT$-symmetric systems \cite{Guo,Longhi,Makris}. One of these is the onset of a phase transition related to the spontaneous symmetry breaking that occurs when the rate between the real and imaginary parts of the refractive index reaches a critical point. As a result, the all-real spectrum undergoes a phase transition to the complex plane. This phase transition has been observed and applied to a number of devices based on $\PT$ optical systems \cite{feng2013,feng2014,science14}.
The interplay between nonlinearity with gain/loss is even a more interesting problem with striking consequences. For example, it has been reported that $\PT$-symmetric nonlinear optical media supports the unexpected possibility of soliton propagation \cite{Christodoulides1}. Further novel properties have been actively explored in recent years \cite{malomed11,konotop12,yang12,segev13,ulf15}.
The developments reported in $\PT$-symmetric optical structures have also triggered further studies
to include non-$\PT$-symmetric complex potential functions widening
the potentiality of $\PT$ symmetry
features such as a real spectra as well a phase transition ~\cite{Cannata1998,Miri2013,Tsoy2014, Yang2016}. These potentials contain
free parameters that represent tuning. By tuning the refractive index below a critical value the real spectrum undergoes a phase transition
to the complex plane. By adding a real
part with an adjustable parameter
one opens a tunable
possibility with ranges that admits real spectra and also a phase transition to the complex phase.
Nonlinear Kerr systems have been studied in a non-$\PT$-symmetric
potential and the continuous families of nonlinear modes reported
there are attributed to a conserved quantity, in the sense that it
does not vary with the transverse spatial coordinate $x$ \cite{npt_solitons} and thus, reflects a content of
symmetry in the system. Recently, we have studied the existence and stability of solitons in quadratic nonlinear unidimensional media
under the influence of a one-parameter family asymmetric class of potentials \cite{Yang2016} and found families of stable solitons \cite{moreira16} bifurcating from the fundamental mode as well as
from the second harmonic. However, to the best of our knowledge, higher nonlinearities in non-$\PT$-symmetric systems have
not been investigated. Materials with high nonlinear coefficients
such as semiconductor doped glasses and organic polymers, cannot be described by a Kerr-type nonlinearity \cite{gatz}, as in these
materials, the saturation of the nonlinear refractive index occurs at moderately high pulse
intensities so that one must include the effect of higher order nonlinearities. Furthermore, theoretical and experimental results have indicated that a saturable nonlinear medium is suitable for the propagation and control of spatial solitons. As a notorious example, photorefractive optical solitons which can be induced at relatively low power formation are quite important for applications in all-optical switching and signal processing \cite{Segev}.
Considering the generalization of the class of functions $V(x)$
(presented in \cite{Yang2016}) that
admit stable nonlinear modes, in the present work we study saturable nonlinear optical waveguides described by a class of tunable complex potentials which are not $\PT$-symmetric. The basic motivation to study
this type of tunable
potential lies on the fact that this additional asymmetry parameter
widens the scope of non-Hermitian Hamiltonians with real spectra,
besides the effects that a saturable nonlinearity may reveal,
paving the way for new metamaterials with designed properties.
We find that this class of tunable asymmetric potentials
admits stable localized modes whose properties are quite modified by
the saturable nonlinearity. The properties of nonlinear modes are investigated, revealing interesting results on the influence of the
asymmetry on the existence and stability of localized solutions which
we now, proceed to describe: in Sec. II we present the model and
statement of the problem. Sec. III presents briefly the linear
properties of the the complex potential. Section IV is devoted to the existence of stable solitons while in section V their stability and dynamic are investigated. Finally, in Sec. VI we conclude and
discuss the results obtained.
\section{Theoretical Model}
Let us begin by writing the propagation equation of an optical field $u(\xi,\zeta)$ through a diffractive nonlinear saturable medium:
\begin{equation}
i\frac{\partial u}{\partial\zeta}+\frac{\partial^{2}u}%
{\partial\xi^{2}}+\left[ g^2+\alpha g+i g_\xi \right]
u
+\sigma \frac{|u|^2}{1+\gamma|u|^2}u=0
\label{final}
\end{equation}
$u$ being the dimensionless amplitude of the field while $\xi$ and $\zeta$ are, respectively, dimensionless transverse and propagation coordinates, scaled to the characteristic size of a localized
modulation of the refractive index which is fully described by a real function $g\equiv g(\xi)$.
The real part being $g^2+\alpha g$, where $\alpha$ is a free parameter which plays the role of a relative strength of the
real part of the potential to the imaginary part, which is defined as: $g_{\xi}\equiv\frac{dg}{d\xi}$. This adjustable parameter $\alpha$, widens the scope of non-Hermitian Hamiltonians with real spectra and may allow the breaking of the real spectrum \cite{Yang2016}. Fig. \ref{fig:FF_phasebreak} (top right panel) shows the variation of the real part $g^2 + \alpha g$ as a function of $\alpha$ illustrating its effect on the amplitude of the real part of the potential function. The last term
in equation ($\ref{final}$) represents a saturable third order nonlinear material with saturation parameter $\gamma$, with $\gamma=0$ representing the usual Kerr nonlinearity without saturation. The nonlinearity may be of a self-focusing $(\sigma > 0)$ or a self-defocusing type, ($\sigma < 0$).
We look
for localized solutions of ($\ref{final}$) in the form,
\begin{equation}
u\left( \xi,\zeta\right) =w\left( \xi\right) e^{ib\zeta}, ,\label{statio}%
\end{equation}
where $b$ is the soliton propagation constant and $w(\xi)$ is the solution of
\begin{equation}
\frac{d^{2}w}{d\xi^{2}}+\left[ g^2+\alpha g+i g_\xi -b\right] w+\sigma \frac{|w|^2}{1+\gamma |w|^2}w,
\label{stat}
\end{equation}
subject to zero boundary conditions $ w\left( \xi\right)\to 0 $ as $|\xi|\to \infty$.
\section{Linear properties of the asymmetric localized potential}
Previous work \cite{Yang2016} has shown that, similarly to the $\PT$-symmetric case, the complex potential
\begin{equation}
G(\xi,\alpha)=g^2+\alpha g +ig_\xi
\end{equation}
when substituted into the linear version of equation (\ref{stat}), that is,
\begin{equation}
Lw=b_1w,
\label{npt_linear}
\end{equation}
where $b_1$ is the linear eigenvalue and the operator $L$ is defined as
\begin{equation}
L=\frac{d^2}{d\xi^2}+G,
\label{lineareqs}
\end{equation}
\noindent depending on the value of $\alpha$, may exhibit all-real spectrum. Furthermore it was proved, for sufficiently mild conditions, that $\alpha=0$ always guarantees a real spectrum for (\ref{lineareqs}). Let us now briefly describe the
recently reported \cite{moreira16} linear properties of (\ref{lineareqs}) when the potential $G$ is generated from the following function:
\begin{equation}
g=V\sech{\left[1+\beta\left(\frac{1+\tanh(\xi)}{2}\right)\right]\xi},
\label{gform}
\end{equation}
where the constant $V$ is the amplitude and the constant $\beta$ is its degree of asymmetry. The function $g$ is continuously differentiable, has a global maximum at $\xi=0$ and is asymmetric for $\beta\neq 0$, as is illustrated in Fig. \ref{fig:FF_phasebreak}. The function $g$ decays as $e^{\xi}$ when $\xi\to-\infty$ and $e^{-\left|1+\beta\right|\xi}$ when $\xi\to\infty$. In Fig. \ref{fig:FF_phasebreak} we show the spectrum of (\ref{lineareqs}) using (\ref{gform}) for different values of $\alpha$ and a fixed $\beta=-0.5$.
\begin{figure}[!htb]
\begin{center}
\scalebox{.43} {\includegraphics{W_scan_V05_ds-05_insept_b1.eps}}
\end{center}
\caption{Upper left panel shows the eigenvalues of (\ref{lineareqs}) with $V=0.5$, $\beta=-0.5$ as a function of the parameter $\alpha$. Upper right panel shows the effect of the constant $\alpha$ in the shape of the real part of the potential $G$. Lower panels show the eigenvalues before ($\alpha=-0.18$) and after ($\alpha=-0.2$) the real spectrum breaks into a complex phase. Note the appearance of a complex pair of eigenvalues due to the collision of two real valued eigenvalues at $b=0.013$ where $\alpha_{cr}=-0.187$ (not shown). The spectrum contains an imaginary pair for $\alpha<\alpha_{cr}$.}
\label{fig:FF_phasebreak}
\end{figure}
A phase transition to the complex plane is found for
$\alpha<\alpha_{cr} = -0.187$, where a pair of complex
eigenvalues appear due to the collision of two localized eigenvalues at $b_1=0.013$ (See Fig.\ref{fig:FF_phasebreak}). Note that for $\alpha>0$ there is no phase transition, instead, there is the appearance of additional localized modes due to the higher amplitude of $G$. Here, in contrast with the symmetric system for which $\beta=0$, there are no degenerated discrete eigenvalues as there is no symmetry to induce degeneracy. Instead, as $\alpha$ changes continuously, a complex
phase may appear only due to the collision of simple discrete
modes as illustrated in Fig. \ref{fig:FF_phasebreak}. Albeit shown
here for the specific case of $\beta=-0.5$, different values of
$\beta$ are found to exhibit the same behavior with respect to
the breaking of the real valued phase of the spectrum. The effect of different values of the asymmetry constant $\beta$ on the
eigenvalues $b_1$ is shown in Fig. \ref{fig:beta_spectrum}. In
subsequent results were considered the values $V=\pm\sqrt{0.5}$.
Note in (\ref{gform}) that this means that the absolute value of
the amplitude of $g$ is fixed, only the decay rate of the $\xi>0$
region will be allowed to change. This affects the values of
$b_1$: as $\beta$ decreases, the area of $g$ becomes larger and, consequently, the number of localized modes increases. Clearly if $\beta=-1$, then
\begin{subequations}
\begin{equation}
\lim_{\xi\to -\infty}G(\xi) =0,
\end{equation}
\begin{equation}
\lim_{\xi\to \infty}G(\xi) =V^2.
\end{equation}
\end{subequations}
\begin{figure}[!htb]
\begin{center}
\scalebox{.55} {\includegraphics{beta_spectrum.eps}}
\end{center}
\caption{Discrete spectrum of $b_1$ as a function of $\beta$. Note the appearance of a continuous band at $\beta=-1$. The values chosen for the potential are $V=\sqrt{0.5}$ and $\alpha=0$.}
\label{fig:beta_spectrum}
\end{figure}
This is an interesting situation as for $\beta=-1$ the potential $G$ becomes very similar to a Heaviside step function with some localized imaginary part while the spectrum approaches a continuous band defined by $b_1\leq V^2$.
\section{Nonlinear localized modes and linear stability analysis}
Let us now, investigate the existence and stability of the localized solutions of (\ref{stat}). From now on we restrict our considerations mainly to solutions bifurcating from the linear limit, which is understood as a limit defined as $w \to 0$, $b$ tends to a finite eigenmode of the linear problem i.e., $b \to b_1$. The total energy $P$ of a localized solution is defined by:
\begin{equation}
P=\int_{\infty}^{\infty}\left|w\right|^{2}d\xi.
\end{equation}
Let us now perturb the solution $u(\xi,\zeta)$ by writing:
\begin{equation}
u(\xi,\zeta)=\left( w(\xi)+p_{+}(\xi)e^{-i\lambda\zeta
}+\overline{p}_{-}(\xi)e^{i\overline{\lambda}\zeta}\right) e^{ib\zeta
},
\label{smallperturbations}
\end{equation}
where, $|p_{\pm}(\xi)|\ll 1$, represents small perturbations. Inserting (\ref{smallperturbations}) into (\ref{final}) and collecting linear terms in $p_{\pm}(\xi)$ one ends up with an eigenvalue problem given by:
\begin{subequations}
\begin{equation}
L_s\left(
\begin{array}
[c]{c}%
p_{2+}\\
p_{-}
\end{array}
\right) =\lambda\left(
\begin{array}
[c]{c}%
p_{+}\\
p_{-}
\end{array}
\right),
\end{equation}
\begin{equation}
L_s=\left(
\begin{array}
[c]{cc}%
L-b+\sigma\frac{|w|^2(\gamma|w|^2+2)}{(1+\gamma|w|^2)^2}& \sigma\frac{w^2}{(1+\gamma|w|^2)^2} \\
-\sigma\frac{\overline{w}^2}{(1+\gamma|w|^2)^2} &-\overline{L}+b-\sigma\frac{|w|^2(\gamma|w|^2+2)}{(1+\gamma|w|^2)^2}
\end{array}
\right)
\end{equation}
\label{linears}%
\end{subequations}
so that, whenever an eigenvalue $\lambda$ with $\text{Im}(\lambda)>0$ occurs the solution is linearly unstable.
%We have found branches bifurcating from all discrete values of $b$ in the case $V=\pm\sqrt{0.5}$.
\subsection{Effects of saturation parameter $\gamma$}
Investigating the effect of the saturation parameter $\gamma$ we have found that all investigated branches with $\gamma>0$ exhibit a vertical asymptote in the
$(b,P)$ plane, suppressing the domain of existence of modes, as can be seen in Fig. \ref{fig:gamma_branchesV05sigma1} and Fig. \ref{fig:gamma_branchesV05sigma-1}.
\begin{figure}[!htb]
\begin{center}
\scalebox{.6} {\includegraphics{branches_V05_ds-05_c0_sigma1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\beta=-0.5$, $\sigma=1$ and different values of the saturation constant $\gamma$. Dashed lines represents the asymptotes $b_1+\sigma/\gamma$. Thick lines represent linearly stable solutions and lines represent unstable solutions.}
\label{fig:gamma_branchesV05sigma1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.67} {\includegraphics{branches_V05_ds-05_c0_sigma-1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\beta=-0.5$, $\sigma=-1$ and various values of the saturation constant $\gamma$. Thick lines represent linearly stable solutions and lines represent unstable solutions.}
\label{fig:gamma_branchesV05sigma-1}
\end{figure}
This is expected since for solutions with very high intensities one has,
\begin{equation}
\lim_{|w(\xi)|^2\to\infty}{\sigma\frac{|w(\xi)|^2}{1+\gamma|w(\xi)|^2}}=\frac{\sigma}{\gamma}
\label{b_limit}
\end{equation}
Now, inserting (\ref{b_limit}) in (\ref{stat}) we obtain
\begin{equation}
\frac{d^{2}w}{d\xi^{2}}+\left[ g^2+\alpha g+i g_\xi -b+\frac{\sigma}{\gamma}\right]w,
\end{equation}
which, of course the same as (\ref{npt_linear})
\begin{equation}
Lw=(b-\frac{\sigma}{\gamma})w
\end{equation}
and has a localized mode at $b=b_1+\frac{\sigma}{\gamma}$, where $b_1$ is a discrete value of (\ref{npt_linear}). In other words, as the amplitude of the soliton increases its profile becomes similar to the shape of the linear eigenmode it bifurcated in the limit $w\to 0$.
\begin{figure}[!htb]
\begin{center}
\scalebox{.59} {\includegraphics{branches_V-05_ds-05_c0_sigma1_final1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\beta=-0.5$, $\sigma=1$ and for various values of saturation constant $\gamma$. Right panel shows an expanded view of the region where the transition between stable and unstable solutions occurs. Thick lines represent linearly stable solutions and lines represent unstable solutions.}
\label{fig:gamma_branchesV-05sigma1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.65} {\includegraphics{stab_P_lambda_gamma.eps}}
\end{center}
\caption{Maximum value of imaginary part of (\ref{linears}) as a function of the total energy of localized modes $P$ for several values of $\gamma$. Note that all branches are unstable but $\max \{\text{Im}\lambda\}\to \lambda_{\gamma}$ as $P\to \infty$. The higher the value of $\gamma$, the lower the valuer $\lambda_\gamma$. Parameters are $V=-\sqrt{0.5}$, $\beta=-0.5$, $\sigma=1$.}
\label{fig:stab_P_lambda_gamma}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.65} {\includegraphics{branches_V-05_ds-05_c0_sigma-1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\beta=-0.5$, $\sigma=-1$ and different values of saturation constant $\gamma$. All solutions are stable.}
\label{fig:branches_V-05_ds-05_c0_sigma-1}
\end{figure}
In fact, extensive numerical calculations of branches show that the fundamental branches exists in the region $b\in \left(b_1,b_1+\gamma^{-1}\right)$ for the $\sigma=1$ case and $b\in\left(b_1-\gamma^{-1},b_1\right)$ when $\sigma=-1$. Interestingly enough as, using the fact that the branch must be localized within the region between $b_1$ and $b_1+\frac{\sigma}{\gamma}$, one realizes that the $\sigma=1$($\sigma=-1$) branch bifurcates to the right(left).
In what concerns stability, we have investigated two main potential functions, i.e., $G(\xi)$ and $G^*(\xi)$ which corresponds to the cases $V=\sqrt{0.5}$ and $V=-\sqrt{0.5}$ respectively if one also substitutes $\alpha\to -\alpha$. One can note that if $w$ is a solution of (\ref{stat}) with $G$ then clearly $w^*$ is a solution of (\ref{stat}) with $G^*$. Both solutions clearly have the same $P$. Despite no difference in the $(b,P)-$curve, the stability scenario of the solutions is completely different in each case, as we demonstrate below. First we choose $V=\sqrt{0.5}$. In both cases, $\sigma=\pm 1$, no linearly unstable solutions were obtained, at least for moderately high amplitude solutions with $P\sim 10$ as is displayed in Fig.\ref{fig:gamma_branchesV05sigma1} and Fig. \ref{fig:gamma_branchesV05sigma-1}.
With respect to the $V=-\sqrt{0.5}$ case, we have found only unstable solutions in the $\sigma=1$ case as depicted in Fig. \ref{fig:gamma_branchesV-05sigma1}. The instability here appears due to an asymmetric internal mode that bifurcates from the continuous spectrum of (\ref{linears}), at $\lambda=\pm b$. This internal mode appears as soon as $P>0$, thus there are no stable solutions because this mode moves to the $\text{Im}\lambda>0$ region. Note that this instability rate is quite low when $|b-b_1|\ll 1$. The Fig. \ref{fig:stab_P_lambda_gamma} clearly shows that branches with higher values of $\gamma$ tend to present a lower degree of instability in this case.
In the $V=-\sqrt{0.5}$ case we have found stable solutions only in the case of the defocusing nonlinearity, $\sigma=-1$ as is shown in Fig.~\ref{fig:branches_V-05_ds-05_c0_sigma-1}.
\subsection{Effects of asymetry parameter $\beta$}
The effect of different values of the asymmetry constant $\beta$ has also been investigated. We have studied all four combinations of $V=\pm\sqrt{0.5}$ and $\sigma=\pm 1$. See in Fig. \ref{fig:beta_branches_sigma1} and Fig. \ref{fig:branches_V05_sat10_c0_sigma-1} the case $V=\sqrt{0.5}$. The effect of $\beta$ in both $\sigma=\pm 1$ cases is to shift $b_1$ closer to $b_1=0.5$ as expected, with more discrete modes appearing as $\beta\to -1$ (See Fig.~\ref{fig:beta_spectrum}).
\begin{figure}[!htb]
\begin{center}
\scalebox{.6} {\includegraphics{branches_V05_sat10_c0_sigma1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\beta=-0.5$, $\sigma=1$ and different values of saturation constant $\gamma$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:beta_branches_sigma1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.6} {\includegraphics{branches_V05_sat10_c0_sigma-1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\gamma=10$, $\sigma=-1$ and several values of saturation constant $\gamma$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V05_sat10_c0_sigma-1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V-05_sat10_c0_sigma-1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\gamma=10$, $\sigma=-1$ and various values of asymmetry constant $\beta$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V-05_sat10_c0_sigma-1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V-05_sat10_c0_sigma1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\gamma=10$, $\sigma=1$ and different values of asymmetry constant $\beta$. Thick lines(lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V-05_sat10_c0_sigma1}
\end{figure}
The vertical asymptote $b=b_1+\sigma/\gamma$ is still present in the branches. Regarding stability, no changes were observed: all solutions were found to be stable in the $V=\sqrt{0.5}$ case as illustrated in Fig.~\ref{fig:beta_branches_sigma1} and Fig.~\ref{fig:branches_V05_sat10_c0_sigma-1}. Solutions with $V=-\sqrt{0.5}$ are stable for $\sigma=-1$ (See Fig.~\ref{fig:branches_V-05_sat10_c0_sigma-1}) and all solutions are unstable in the $\sigma=1$ case (See Fig.~\ref{fig:branches_V-05_sat10_c0_sigma1}).
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{stab_P_lambda.eps}}
\end{center}
\caption{Maximum value of the imaginary part of (\ref{linears}) as a function of total energy of localized modes $P$. Note that all
$\beta<0$ branches are unstable but $\max \{\text{Im}\lambda\}\to 0$ as $\beta\to 0$. Parameters are $V=-\sqrt{0.5}$, $\gamma=10$,
$\sigma=1$.}
\label{fig:stab_P_lambda}
\end{figure}
Note that in both $\sigma=\pm 1$ cases with $V=-\sqrt{0.5}$ there are stable solutions only if $\beta=0$, i.e., the $\PT$-symmetric case. In Fig. \ref{fig:stab_P_lambda} we can see that in fact $\max \{\text{Im}\lambda\}\to 0$ as $\beta\to 0$.
\subsection{Investigation of gain and loss parameter $\alpha$}
Here we fix $\beta=-0.5$, $\gamma=10$ and change $\alpha$ to investigate its effects on the existence and stability of fundamental branches. The first case to be investigated is the one with $V=\sqrt{0.5}$ and $\sigma=1$ and we turn to Fig.~\ref{fig:branches_V05_sat10_ds-05_sigma1}, from where one may conclude that all observed solutions are stable. The second case with $V=\sqrt{0.5}$ and $\sigma=-1$ shows that an unstable region may exist above some threshold in $P$ if $\alpha<0$ is close enough to $\alpha_{cr}$. The stable region becomes smaller as $\alpha$ gets closer to $\alpha_{cr}$, eventually disappearing along with the branch itself when $\alpha\to\alpha_{cr}$. Next, in the $V=-\sqrt{0.5}$ and $\sigma=1$ case we note that there are no stable solutions. Finally, in the $V=-\sqrt{0.5}$ and $\sigma=-1$, all solutions are stable.
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V05_sat10_ds-05_sigma1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\gamma=10$, $\sigma=1$ and various values of $\alpha$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V05_sat10_ds-05_sigma1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V05_sat10_ds-05_sigma-1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=\sqrt{0.5}$, $\gamma=10$, $\sigma=-1$ and several values of $\alpha$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V05_sat10_ds-05_sigma-1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V-05_sat10_ds-05_sigma1.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\gamma=10$, $\sigma=1$ and different values of $\alpha$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V-05_sat10_ds-05_sigma1}
\end{figure}
\begin{figure}[!htb]
\begin{center}
\scalebox{.7} {\includegraphics{branches_V-05_sat10_ds-05_sigma-1_final.eps}}
\end{center}
\caption{Total energy $P$ as a function of $b$ with $V=-\sqrt{0.5}$, $\gamma=10$, $\sigma=1$ and different values of asymmetry constant $\beta$. Thick lines (lines) represent linearly stable (unstable) solutions.}
\label{fig:branches_V-05_sat10_ds-05_sigma-1}
\end{figure}
\section{Dynamics of localized solutions}
To corroborate the results described in the last section, in this section we proceed to make comparisons between the model of perturbations (\ref{linears}) and direct propagation of (\ref{stat}) using the solutions found as initial conditions in (\ref{final}). In all propagations we have added a $1\%$ of noise to the amplitude at $\zeta=0$. Here our objective is to show that the linear stability analysis predicts quite well the evolution of localized modes under weak perturbations.
\begin{figure}[ht]
\begin{center}
\scalebox{.75} {\includegraphics{prop_V05_sat10_ds-05_c0_sigma1_highpower.eps}}
\end{center}
\caption{(Color online)(Left panel) Intensity $|u|^2$ during propagation of a high power $P=10$ stable soliton with $b=0.23$. (Right panel) Corresponding eigenvalues of (\ref{linears}). Parameters are $V=\sqrt{0.5},\beta=-0.5,\alpha=0,\gamma=10$ and $\sigma=1$.}%
\label{fig:prop_V05_sat10_ds-05_c0_sigma1_highpower}%
\end{figure}
\begin{figure}[ht]
\begin{center}
\scalebox{.83} {\includegraphics{prop_beta_sigma1.eps}}
\end{center}
\caption{(Color online)(Left panels) Intensities $|u|^2$ during propagation of stable solutions of branches with different values of assymetry constant $\beta$. Parameters are $V=\sqrt{0.5},\alpha=0,\gamma=10$ and $\sigma=1$.}%
\label{fig:prop_beta_sigma1}%
\end{figure}
\begin{figure}[ht]
\begin{center}
\scalebox{.68} {\includegraphics{prop_V05_sat10_ds-05_c-018_sigma1.eps}}
\end{center}
\caption{(Color online)(Left panels) Intensity $|u|^2$ during propagation of a stable solution with $b=0.1$. Parameters are $V=\sqrt{0.5},\beta=-0.5,\alpha=-0.18,\gamma=10$ and $\sigma=1$.}%
\label{fig:prop_V05_sat10_ds-05_c-018_sigma1}%
\end{figure}
The first property that we have observed is the stability of three among four possible combinations of signals of $V$ and $\sigma$, three cases: $V=\sqrt{0.5}$ and $\sigma=\pm1$;$V=-\sqrt{0.5}$ and $\sigma=-1$, exhibit stable solutions as possible to see in Fig.~\ref{fig:prop_V05_sat10_ds-05_c0_sigma1_highpower}, Fig.~\ref{fig:prop_beta_sigma1} and Fig.~\ref{fig:prop_V05_sat10_ds-05_c-018_sigma1} for the $V=\sqrt{0.5}$ and $\sigma=1$ case. In all figures of dynamics we used $\gamma=10$ as it does not change the existence or absence of stable solutions. Note that discrete modes of stable solitons with $\beta\neq 0$ go into the negative imaginary part plane, differently from the $\beta=0$ case, where discrete modes of stable solitons must have $\text{Im}\lambda=0$ as possible to see in Fig.~\ref{fig:prop_beta_sigma1} . In Fig. \ref{fig:prop_V05_sat10_ds-05_c-018_sigma1} we show the evolution of a stable solution with $\alpha=-0.18$, close but below the phase transition of the real spectrum.
An example of the dynamics of an unstable solution for the case $V=\sqrt{0.5}$, $\gamma=10$ and $\sigma=-1$ is shown in Fig.~\ref{fig:prop_V05_sat10_ds-09_c0_sigma-1} for the $\alpha=0$ case with $\beta=-0.9$. In Fig.~\ref{fig:prop_V05_sat10_ds-05_c-014_sigma-1} we show two examples of unstable solutions in the branch with $\alpha=-0.14$ $\beta=-0.5$. Note that while $\beta$ does not change by itself the stability properties of the solutions when all other parameters are fixed, the parameter $\alpha<0$ may in fact introduce instability in the system, both of oscillatory (complex $\lambda$) and purely exponentially decaying (purely complex $\lambda$) types, as displayed in Fig.~\ref{fig:prop_V05_sat10_ds-05_c-014_sigma-1}.
\begin{comment}
\begin{figure}[ht]
\begin{center}
\scalebox{.85} {\includegraphics{prop_V05_sat10_ds-05_c-01_sigma-1.eps}}
\end{center}
\caption{(Color online)(Left panels) Intensities $|u|^2$ during propagation of stable solutions. (Right panels) Corresponding eigenvalues of (\ref{linears}) Note the that de linear stability model correctly predicts the onset of a oscillatory instability in the solution with $b=0.0153$. Parameters are $V=\sqrt{0.5},\alpha=-0.1,\beta=-0.5,\gamma=10$ and $\sigma=-1$.}%
\label{fig:prop_V05_sat10_ds-05_c-01_sigma-1}%
\end{figure}
\end{comment}
\begin{figure}[ht]
\begin{center}
\scalebox{.73} {\includegraphics{prop_V05_sat10_ds-09_c0_sigma-1.eps}}
\end{center}
\caption{(Color online)(Left panel) Intensity $|u|^2$ during propagation for an unstable solution. Parameters are $V=\sqrt{0.5},\beta=-0.9,\alpha=0,\gamma=10$ and $\sigma=-1$.}%
\label{fig:prop_V05_sat10_ds-09_c0_sigma-1}%
\end{figure}
\begin{figure}[ht]
\begin{center}
\scalebox{.75} {\includegraphics{prop_V05_sat10_ds-05_c-014_sigma-1.eps}}
\end{center}
\caption{(Color online)(Left panels) A Intensities $|u|^2$ during propagation of unstable solutions. (Right panels) Corresponding eigenvalues of (\ref{linears}) Note the that de linear stability model correctly predicts the onset of a oscillatory instability in the solution with $b=0.0153$ and the exponential decay in the case $b=0.0006367$. Parameters are $V=\sqrt{0.5},\beta=-0.5,\alpha=-0.14,\gamma=10$ and $\sigma=-1$.}%
\label{fig:prop_V05_sat10_ds-05_c-014_sigma-1}%
\end{figure}
Dynamics of propagation in case $V=-\sqrt{0.5}$ and $\sigma=1$ are shown in both cases, i.e., unstable (Fig.\ref{fig:prop_V-05_sat10_ds-05_c0_sigma1} ) and stable (Fig.\ref{fig:prop_V-05_sat10_ds-05_c0_sigma-1_final}).
\begin{figure}[ht]
\begin{center}
%\scalebox{.75} {\includegraphics{prop_V-05_sat10_ds-05_c0_sigma1.eps}}
\scalebox{.75} {\includegraphics{prop_V-05_sat10_ds-05_c0_sigma1_final.eps}}
\end{center}
\caption{(Color online)(Left panels) Intensities $|u|^2$ during propagation for unstable solutions. (Right panels) Corresponding eigenvalues of (\ref{linears}). Note the that de linear stability model correctly predicts the onset of a oscillatory instability. Also note that the instability develops quite slowly. Parameters are $V=-\sqrt{0.5},\beta=-0.5,\alpha=0,\gamma=10$ and $\sigma=1$.}%
\label{fig:prop_V-05_sat10_ds-05_c0_sigma1}%
\end{figure}
Finally in Fig.\ref{fig:prop_V-05_sat10_ds-05_c0_sigma-1} we show the dynamics of a stable solution in the case $V=-\sqrt{0.5}$ and $\sigma=-1$.
\begin{figure}[ht]
\begin{center}
\scalebox{.73} {\includegraphics{prop_V-05_sat10_ds-05_c0_sigma-1.eps}}
\end{center}
\caption{(Color online)(Left panel) Intensity $|u|^2$ during propagation for a stable solution. Parameters are $V=-\sqrt{0.5},\beta=-0.5,\alpha=0,\gamma=10$ and $\sigma=-1$.}%
\label{fig:prop_V-05_sat10_ds-05_c0_sigma-1}%
\end{figure}
\begin{figure}[ht]
\begin{center}
\scalebox{.73} {\includegraphics{prop_V-05_sat10_ds-05_c0_sigma-1_final.eps}}
\end{center}
\caption{(Color online)(Left panel) Intensity $|u|^2$ during propagation for a high power, $P=10$, stable solution with $b=0.0412$. Parameters are $V=-\sqrt{0.5},\beta=-0.5,\alpha=0,\gamma=10$ and $\sigma=-1$.}%
\label{fig:prop_V-05_sat10_ds-05_c0_sigma-1_final}%
\end{figure}
Looking at the stability spectrum of (\ref{linears}) we can see that the role of the sign of $V$ has in respect to stability of branches can be explained: First we see that non-$\PT$-symmetric cases, i.e., $\beta\neq 0$, do not possess the symmetry $\lambda=\lambda^*$ as expected. So one can have discrete eigenvalues $\lambda$ in either $\text{Im}\lambda>0$ or $\text{Im}\lambda<0$ regions. Now, if one investigate a branch where the potential is substituted for $G\to G^*$ the spectrum of (\ref{linears}) changes to $\lambda\to\lambda^*$ as all terms other terms other than $L$ are real in $L_s$. So whenever an stable solution exists with some $\text{Im}\lambda<0$ for a given $G$ there is a unstable solution in the branch of the system with the potential $G^*$. There are cases where changing $G$ for $G^*$ does not change the stability of a function. This happens when a stable solution only has $\text{Im}\lambda=0$ as possible to see in Fig. \ref{fig:prop_V05_sat10_ds-05_c0_sigma1_highpower} and Fig~\ref{fig:prop_V-05_sat10_ds-05_c0_sigma-1_final} for examples.
%\begin{figure}[ht]
%\begin{center}
%\scalebox{.63} {\includegraphics{prop_V05_ds-05c0_FF.eps}}
%\end{center}
%\caption{(Color online)(Left panels) Intensities $|u_{1}|^2$ during propagation for stable( upper and lower panels) and an unstable (middle panel). (Right panels) Corresponding eigenvalues of (\ref{linears}) Note the that de linear stability model correctly predicts the onset of a oscillatory instability in the solution with $b=0.14$. All solutions pertain to the $b$ branch. Parameters are $\alpha=0$ and $\beta=-0.5$}%
%\label{fig:prop_V05_ds-05c0_FF}%
%\end{figure}
\section{Conclusions}
In conclusion, we have studied the existence and stability properties of soliton families in media with a saturable nonlinearity embedded in a non-$\PT$-symmetric complex refractive index. We
have found the that such media does support the existence of asymmetric solitons in both self-focusing as well self-defocusing case. Furthermore,
all branches of solitons found to bifurcate from the linear limit in case the saturation parameter $\gamma \neq 0$, exhibit a vertical asymptote in the power-eigenvalue diagram, revealing that $\gamma$ restricts severely
the range where localized modes that are supported, in comparison
with a Kerr nonlinearity where there is no restriction. Also, the $b$ asymptotic value for large powers is uniquely determined by the saturation
parameter, and larger saturation parameters lead to greater reduction in
the mode range. This result is similar to the result reported in ${\cal PT}$- optical lattices \cite{Hu-Hu}. Furthermore, as the amplitude of the soliton increases its profile becomes similar to the shape of the linear eigenmode from where it was originated. In case of positive $V$ we find stable solitons in both media, i.e., under the effect of self-focusing or self-defocusing. In cases where $V$ assumes negative values we find that, as the amplitude of the soliton depends on $V^2$ nothing changes in the spectrum. However, when it comes to the stability properties these are complete altered and all solutions become unstable in the
self-focusing media, i. e. for $\sigma = 1$. Variations on the value of $\alpha$ do not change the overall picture, except for the locations where the nonlinear modes are created. All the analytic results have been
confronted with numerical simulations displaying the propagation of
stable and unstable nonlinear modes with a remarkable agreement.
We hope that our investigation on waveguides, combining saturable nonlinearity and asymmetric complex potentials reveals itself to be
useful for applications
in optics widening the class of metamaterials that may be tailored
to control light propagation in waveguides and thus, providing novel techniques.
\acknowledgments
The authors grateful acknowledge the partial financial support from the Brazilian agencies, CNPq and Capes.
\begin{thebibliography}{99}
\bibitem{Bender1}
C. M. Bender and S. Boettcher, Phys.Rev.Lett. {\bf 80}, 5243 (1998);%A. Mostafazadeh %Pseudo-Hermiticity for a class of nondiagonalizable Hamiltonians. Journal of Mathematical Physics,
%{\bf 43}, 6343 (2002)
\bibitem{Ruter2010} C. E. Ruter, K. G. Makris, R. El-Ganainy, D. N. Christodoulides, M. Segev, and D. Kip, Nat. Phys. {\bf 6}, 192 (2010).
\bibitem{peschel} A. Regensburger, C. Bersch, M. A. Miri, G. Onishchukov, D. N. Christodoulides, and U. Peschel, Nature (London) {\bf 488}, 167
(2012).
\bibitem{peng} B. Peng, S. O ̈zdemir, F. Lei, F. Monifi, M. Gian- freda, G. Long, S. Fan, F. Nori, C.M. Bender, and L. Yang, Nat. Phys. {\bf10}, 394 (2014).
\bibitem{Guo} A. Guo, G. J. Salamo, D. Duchesne, R. Morandotti, M. Volatier-Ravat, V. Aimez, G. A. Siviloglou and D. N.
Christodoulides Phys. Rev. Lett. {\bf103}, 093902 (2009).
\bibitem{Longhi} S. Longhi, Phys. Rev. Lett. {\bf105}, 013903 (2010).
\bibitem{Makris} K. G. Makris, R. El-Ganainy, D. N. Christodoulides, and Z. H. Musslimani, Phys. Rev. Lett. 100, 103904 (2008).
\bibitem{feng2013} L. Feng, Y.L. Xu, W.S. Fegadolli, M.H. Lu, J.E.B. Oliveira, V.R. Almeida, Y.F. Chen,
and A. Scherer,Nature Materials {\bf12}, 108 (2013).
\bibitem{feng2014} L. Feng, Z.J. Wong, R. Ma, Y. Wang, and X. Zhang, Science {\bf346}, 972 (2014).
\bibitem{science14} H. Hodaei, M.A. Miri, M. Heinrich, D. N. Christodoulides, and M. Khajavikhan, Science {\bf346}, 975 (2014).
\bibitem{Christodoulides1} Z. H. Musslimani, K. G. Makris, R. El-Ganainy, and D. N. Christodoulides
Phys. Rev. Lett. {\bf 100}, 030402, (2008).
\bibitem{malomed11} R. Driben and B. A. Malomed, Opt. Lett. {\bf36}, 4323 (2011).
\bibitem{konotop12} D. A. Zezyulin and V.V. Konotop, Phys. Rev. A {\bf85}, 043840 (2012).
\bibitem{yang12} S. Nixon, Y. Zhu and J. Yang, Opt. Lett. {\bf37}, 4874 (2012).
\bibitem{segev13} Y. Lumer, Y. Plotnik, M. C. Rechtsman, and M. Segev, Phys. Rev. Lett. {\bf111}, 263901 (2013).
\bibitem{ulf15} M. Wimmer, A. Regensburger, M. Miri, C. Bersch, D. N. Christodoulides and U. Peschel, Nature Comm. {\bf6} 7782 (2015).
\bibitem{Cannata1998} F. Cannata, G. Junker, and J. Trost, Phys. Lett. A {\bf 246},219â226 (1998).
\bibitem{Miri2013} M.A. Miri, M. Heinrich, and D. N. Christodoulides, Phys. Rev. A {\bf 87}, 043819 (2013).
\bibitem{Tsoy2014} E.N. Tsoy, I.M. Allayarov and F. Kh. Abdullaev, Opt. Lett. {\bf 39}, 4215â4218 (2014).
\bibitem{Yang2016} S. Nixon and J. Yang, Phys. Rev. A {\bf 93}(3), 031802 (2016).
\bibitem{npt_solitons} E. N. Tsoy, I. M. Allayarov, and F. Kh. Abdullaev, Opt. Lett.
\bibitem{moreira16} F. C. Moreira and S. B. Cavalcanti, Phys. Rev. A {\bf 94}, 043818.
\bibitem{gatz} S. Gatz and J. Herrmann, J. Opt. Soc. Am. B {\bf 8}, 2296-2302 (1991)
\bibitem{Segev} M. Segev, B. Crosignani, A. Yariv and B. Fischer, Phys. Rev. Lett. {\bf 68}, 923 (1992).
\bibitem{Znojil}
M. Znojil, J.Phys. A Math. Gen. {\bf 33}, L61 (2000).
\bibitem{Ahmed}
%Z. Ahmed, Phys.Lett. A {\bf 282}, 343 (2001).
Z. Ahmed, Phys. Lett. A {\bf 282}, 343-348 (2001).
\bibitem{Gaussian} S. Hu, Xuekai Ma, Daquan Lu, Zhenjun Yang, Yizhou Zheng, and Wei Hu %Solitons supported by complex $\PT$--symmetric Gaussian potentials.
Phys. Rev. A, {\bf 84}, 043818 (2011).
\bibitem{defoc} Zhiwei Shi, Xiujuan Jiang, Xing Zhu, and Huagang Li, Phys. Rev. A {\bf 84}, 053855 (2011)
\bibitem{AKOS} F. K. Abdullaev, V. V. Konotop, M. \"Ogren, and M. P. S\o rensen,
%Zeno effect and switching of solitons in nonlinear couplers.
Opt. Lett., {\bf 36}, 4566 (2011).
\bibitem{double_well} H. Cartarius and G. Wunner arXiv:1203.1885v1 (2012).
\bibitem{Moreira_loc} F. C. Moreira, F. Kh. Abdullaev, V. V. Konotop, and A. V. Yulin, Phys. Rev. A {\bf 86}, 053815 (2012).
\textbf{39}, 4215 (2014); V. V. Konotop and D. A. Zezyulin, Opt. Lett. \textbf{39}, 5535 (2014).
\bibitem{Hu-Hu} S. Hu and Wei Hu, Physica B {\bf 429}, 28 (2013).
\begin{comment}
\bibitem{Muga}
A. Ruschaupt, F. Delgado and J. G. Muga, J. Phys. A {\bf 38} L171 (2005).
\bibitem{Exp1}
A. Guo, G. J. Salamo, D.Duchesne,R. Morandotti, M. Volatier-Ravat, V. Aimez, G.A. Siviloglou, D.N. Christodoulides,
Phys. Rev. Lett., {\bf 103}, 093902 (2009); C. E. R\"uter, K.G.Makris, R. El-Ganainy,D.N. Christodoulides, M. Segev, and D. Kip, Nat. Phys. 6, 192 (2010).
\bibitem{Mostafa_scattering} A. Mostafazadeh,
%Resonance phenomenon related to spectral singularities, complex barrier potential, and resonating %waveguides.
Phys. Rev. A, {\bf 80} 032711 (2009).
\bibitem{Kottos} O. Bendix, R. Fleischmann, T. Kottos, and B. Shapiro,
%Exponentially Fragile $\PT$- Symmetry in Lattices with Localized Eigenmodes.
Phys. Rev. Lett. {\bf 103}, (2009)
\bibitem{PT-_sol}
F. Kh. Abdullaev, V.V. Konotop, M. Salerno, and A. V. Yulin, Phys. Rev. E {\bf 82}, 056606 (2010);
F. Kh. Abdullaev, Y. V. Kartashov, V.V. Konotop, and D. A. Zezyulin, Phys. Rev. A {\bf 83}, 041805(R) (2011);
Y. He , X. Zhu, D. Mihalache, J. Liu and Z. Chen, Phys. Rev. A {\bf 85}, 013831 (2012);
S. Nixon, L. Ge, and J.Yang, Phys. Rev A {\bf 85}, 023822 (2012);
V. Achilleos, P. G. Kevrekidis, D. J. Frantzeskakis, and R. Carretero-Gonzales, Phys. Rev. A {\bf 86 }, 013808 (2012);
J. Zeng and Y. Lan, Phys.Rev. E {\bf 85}, 047601 (2012); S.V. Suchkov, S.V. Dmitriev, B. A. Malomed, and Yu. S. Kivshar, Phys. Rev. A {\bf 85}, 033825 (2012).
%\bibitem{Suchkov}
%S. Suchkov et al. Phys. Rev. A {\bf 85}, 033825 (2012).
\bibitem{Clausen1}
C.B. Clausen, J.P. Torres, and L. Torner, Phys. Lett. A {\bf 249}, 455 (1998).
\bibitem{Clausen2}
C.B. Clausen and L. Torner, Phys. Rev. Lett. {\bf 81}, 790 (1998).
%\bibitem{Moreira_per} F. C. Moreira, V. V. Konotop, and B. A. Malomed, Phys. Rev. A {\bf 87}, 013832 (2013).
\bibitem{Ahmmed2001} Z. Ahmed, Phys. Lett. A {\bf 282}, 343 (2001)
\bibitem{Yang2015}S. Nixon and J. Yang, Phys. arXiv \textit{preprint} arXiv:1509.07057 (2015).
Phys. Rev. A 93, 031802(R) â Published 9 March 2016 All-real spectra in optical systems with arbitrary gain and loss distributions
\bibitem{Midya}
B. Midya, B. Roy, R. Roychoudhury, Phys. Lett. A {\bf 374} 2605-2607 (2010).
\bibitem{Landau}
D. Landau, E. M. Lifshitz, \textit{Quantum Mechanics: Non-Relativistic Theory, Volume {\bf 3}}, (Elsevier Science, MA, 1958)
\bibitem{Zezyulin}
D. A. Zezyulin and V.V. Konotop, Phys. Rev. Lett. {\bf 108}, 213906 (2012).
\bibitem{Tsoy}
E.N. Tsoy, S. Tadjimuratov, and F.Kh. Abdullaev, Opt. Commun. {\bf 285}, 3441 (2012).
%\bibitem{Legendre}
%F. W. J. Olver, D. W. Lozier, R. F. Boisvert, C. W. Clark. \textit{NIST Handbook of Mathematical functions}, (Cambridge University Press, 2010)
%\bibitem{Bender_coupling}
%C. M. Bender,and H. F. Jones,
%"Interactions of Hermitian and non-Hermitian Hamiltonians."
%J. Phys. A {\bf 41}, 244006 (2008)
%\bibitem{Bender2}
%C. Bender arXive...
%\bibitem{Znojil}
%M. Znojil, J.Phys. A Math. Gen. {\bf 33}, L61 (2000).
\bibitem{Jacobi}
I.S. Gradshteyn, I.M. Ryzhik, Table of Integrals, Series and Products, 4th ed., Academic Press, New York, 1980, p. 838, (7.375.2).
\end{comment}
\end{thebibliography}
\end{document}
%Therefore,
%investigations on non-Hermitian Hamiltonians besides being important
%from the $\PT$ symmetry theoretical point of view, are also quite important to shed light on dissipation/gain processes in optical structures, so important to light control and usually difficult to deal with, in real systems. Actually,