nlin0007042/qpm6.tex
1: \documentstyle[aps,multicol,psfig]{revtex}
2: 
3: \begin{document}
4: 
5: \title{Solitons in quadratic nonlinear photonic crystals}
6: \author{Joel F.~Corney and Ole Bang}
7: \address{Department of Mathematical Modelling, 
8: Technical University of Denmark, Bldg.~321, 2800 Kgs.~Lyngby, Denmark}
9: 
10: \maketitle 
11: 
12: \begin{abstract}
13: We study solitons in one-dimensional quadratic nonlinear photonic crystals 
14: with modulation of both the linear and nonlinear susceptibilities.  
15: We derive averaged equations that include induced cubic nonlinearities 
16: and numerically find previously unknown soliton families.
17: The inclusion of the induced cubic terms enables us to show that solitons
18: still exist even when the {\em effective quadratic nonlinearity vanishes} 
19: and conventional theory predicts that there can be no soliton. 
20: We demonstrate that both bright and dark forms of these solitons
21: are stable under propagation.
22: \end{abstract}
23: 
24: \pacs{PACS number: 42.65.Ky; 42.65.Jx; 42.65.Tg}
25: 
26: \begin{multicols}{2}
27: 
28: \newlength{\figwidtha}
29: \newlength{\figwidthb}
30: \setlength{\figwidtha}{0.45\linewidth}
31: \setlength{\figwidthb}{0.95\linewidth}
32: 
33: The physics and applications of photonic band-gap (PBG) materials, or 
34: {\em photonic crystals}, have been active topics of research for more 
35: than a decade.
36: The theory of linear photonic crystals is now well understood, and many 
37: of their fundamental properties and technical applications have been 
38: characterized \cite{Sou96}.
39: The next important step in the application of photonic crystals is to 
40: create tunable PBGs.
41: Tunability is possible in linear photonic crystals through, e.g., the 
42: temperature dependence of the refractive index \cite{Leo_etal00} or the 
43: electro-optic effect \cite{BusJoh99}.
44: Ultrafast dynamical tunability of the PBG can be accomplished using
45: nonlinearity, as was first demonstrated with a constant Kerr nonlinearity 
46: \cite{LarHibMizSte90}.
47: 
48: Here we consider {\em quadratic nonlinear photonic crystals} (QNPCs) that have a linear grating (periodic dielectric constant) and/or a nonlinear 
49: grating (periodic second-order or $\chi^{(2)}$ susceptibility). 
50: QNPCs are of interest for all-optical components due to the fast and 
51: strong nonlinearity they can provide through the parametric cascading 
52: effect \cite{SteHagTor96}.  
53: The efficiency of the cascading process depends critically on the 
54: phase mismatch between the fundamental and second-harmonic (SH) waves,
55: but two powerful methods exist that use exactly a periodic photonic 
56: crystal structure to control the mismatch 
57: \cite{ArmBloDucPer62,TanBey73,JasArvLau86,FejMagJunBye92}. 
58: In one method, a QNPC with a linear Bragg grating is used to bend 
59: the dispersion curve near the PBG \cite{TanBey73,NPC_chi2_bragg}. 
60: However, the short period, which is of the order of the optical wavelength, can be 
61: inconvenient.
62: The second scheme, quasi-phase-matching (QPM), controls the phase mismatch 
63: \cite{ArmBloDucPer62,FejMagJunBye92} using a nonlinear grating with a
64: period equal to the comparatively long beat length (typically of the 
65: order of microns). 
66: QPM is also possible with linear gratings, but this is much less 
67: effective \cite{TanBey73,JasArvLau86}.
68: 
69: One of the spectacular manifestations of nonlinearity is the soliton, 
70: a self-localized entity that can propagate unchanged over long 
71: distances.
72: Homogeneous $\chi^{(2)}$ materials support solitons in all dimensions 
73: \cite{Tor98}, and gap solitons exist in QNPCs with a linear Bragg grating 
74: \cite{gap}.
75: In this Letter we focus on the open fundamental problem of whether solitons 
76: exist in 1D QNPCs with {\em both a linear and a nonlinear QPM grating}. 
77: Such a simultaneous linear grating is difficult to avoid when, for example, 
78: creating nonlinear QPM gratings in GaAs/AlAs semiconductors through 
79: quantum-well disordering \cite{Hel00}.
80: 
81: Solitons exist in 1D QNPCs with a nonlinear QPM grating \cite{ClaBanKiv97}, 
82: but a simultaneous linear grating can reduce the effective $\chi^{(2)}$
83: nonlinearity \cite{TanBey73,JasArvLau86,FejMagJunBye92}.
84: Thus the global existence of solitons in such QNPCs is nontrivial.
85: Building on previous findings that nonlinear QPM gratings induce 
86: cubic nonlinearities \cite{ClaBanKiv97}, we find soliton solutions 
87: that are stable under propagation.  
88: The induced cubic terms provide an elegant means of showing that, against 
89: intuition, {\em the QNPC supports stable bright and dark solitons even 
90: when there is no effective $\chi^{(2)}$ nonlinearity}.  
91: This is analogous to the existence of solitons in dispersion-managed 
92: fibres with no average dispersion \cite{DMsol}.
93: 
94: \begin{figure}
95:   \vspace{-8mm}
96:   \centerline{\hbox{\psfig{figure=pap39_newfig1.ps,width=\figwidthb}}}
97:   \vspace{-5mm}
98:   \caption{Geometry of the 1D QNPC in the form of a $\chi^{(2)}$ slab 
99:   waveguide. Black and white domains indicate regions with different 
100:   dielectric constant and $\chi^{(2)}$ coefficient.} 
101:   \label{geometry} 
102: \end{figure}
103: 
104: We consider the interaction of a cw beam (carrier frequency $\omega$) 
105: with its SH, propagating in a lossless 1D QNPC under conditions for type I 
106: second-harmonic generation (SHG), as sketched in Fig.~\ref{geometry}.
107: We assume that the modulation of the refractive index is weak ($\Delta n_j(z)/\bar{n}_j$$\ll$1, where $n_j(z)$=$\bar{n}_j+\Delta n_j(z)$ and $j$ refers to the 
108: frequency $j\omega$), and we consider only gratings for forward QPM.
109: The grating period is then much longer than the optical period, in which 
110: case Bragg reflections can be neglected.  The evolution of the slowly 
111: varying beam envelopes is then described by
112: \cite{ArmBloDucPer62,MenSchTor94}
113: \begin{eqnarray}
114:   i\frac{\partial w}{\partial z} + \frac{1}{2}
115:   \frac{\partial^2 w}{\partial x^2} + \alpha_1(z) w + 
116:   \chi(z)w^*v {\rm e}^{i\beta z} & = & 0 ,\nonumber\\
117:   i\frac{\partial v}{\partial z} + \frac{1}{4}
118:   \frac{\partial^2 v}{\partial x^2} + 2\alpha_2(z) v + 
119:   \chi(z)w^2 {\rm e}^{-i\beta z} & = & 0, 
120:   \label{field_eqns}
121: \end{eqnarray}
122: where $w$=$w(x,z)$ and $v$=$v(x,z)$ are the envelope functions of the 
123: fundamental and SH, respectively.
124: The transverse and propagation coordinates $x$ and $z$ are in units of 
125: the input beam width $x_0$ and the diffraction length $L_d$=$k_1 x_0^2$, 
126: respectively. 
127: The parameter $\beta$=$\Delta k L_d$ is proportional to the mismatch 
128: $\Delta k$=$k_2$-$2k_1$, $k_j$=$j\omega\bar{n}_j/c$ being the average 
129: wavenumber. Thus $\beta$ is positive for normal dispersion and negative 
130: for anomalous dispersion.
131: The normalized refractive index grating is given by $\alpha_j(z)$=$L_d
132: \omega\Delta n_j(z)/c$ and the normalized nonlinear grating by 
133: $\chi(z)$=$L_d\omega d_{\rm eff}(z)/(\bar{n}_1 c)$, where 
134: $d_{\rm eff}$=$\chi^{(2)}/2$ is given in MKS units.
135: The model (\ref{field_eqns}) describes both temporal and spatial solitons
136: \cite{MenSchTor94}.
137: 
138: \begin{figure}
139:   \centerline{\hbox{\psfig{figure=pap39_newfig2.ps,width=\figwidthb}}}
140:   \vspace{-5mm}
141:   \caption{Normalized linear and quadratic nonlinear gratings, 
142:   $\alpha_j(z)$ and $\chi(z)$, with period $2L_0$=$2\pi/|\kappa|$.} 
143:   \label{modulation}
144: \end{figure}
145: 
146: The aim is now to average Eqs.~(\ref{field_eqns}) and derive accurate
147: equations for the average field. To do so we focus on first order QPM 
148: using the conventional square gratings with 50\% duty cycle, shown in 
149: Fig.~\ref{modulation}.  We expand the grating functions in Fourier series:
150: \begin{equation}
151:   \label{FourGrating}
152:   \alpha_j(z) = a_j \sum_n g_n {\rm e}^{in\kappa z} , \;
153:   \chi(z)     = d_0 + d \sum_n g_n {\rm e}^{in\kappa z},
154: \end{equation}
155: where $g_n$=$2s/(i\pi n)$ for $n$ odd and $g_n$=0 for $n$ even, with $s = {\rm sign}(\kappa)$.
156: The gratings drive the system, which means that we may expand the 
157: envelope functions in Fourier series also:
158: \begin{equation}
159:   \label{FourField}
160:   w = \sum_n w_n(z,x) {\rm e}^{in\kappa z} , \quad
161:   v = \sum_n v_n(z,x) {\rm e}^{i(n\kappa - \tilde\beta)z} ,
162: \end{equation}
163: assuming that the coefficients $w_n(z,x)$ and $v_n(z,x)$ vary 
164: slowly in $z$ compared to $\exp(i\kappa z)$.
165: The residual mismatch $\tilde\beta$=$\beta-\kappa$ is ideally zero.  
166: 
167: Three physical length scales are in play: the diffraction length $L_d$, 
168: the coherence length $L_c$, and the grating domain length $L_0$. 
169: In normalized units $L_d$=1, $L_c$=$\pi/|\beta|$, and $L_0$=$\pi/|\kappa|$.
170: We assume a typical QPM grating with a domain length that is much shorter 
171: than the diffraction length, $L_0$$\ll$1. 
172: Furthermore, the grating is of good quality, with the domain length being 
173: close to the coherence length, $L_0$$\simeq$$L_c$, so the residual 
174: mismatch is small, $|\tilde\beta|$$\ll$$|\kappa|$.
175: In this case $|\kappa|$$\gg$1 and we can use perturbation theory with 
176: the small parameter $\epsilon$=$1/|\kappa| \ll 1$. 
177: 
178: Following the approach of Ref.~\cite{ClaBanKiv97}, we insert the Fourier
179: expansions (\ref{FourGrating}) and (\ref{FourField}) into the dynamical 
180: equations and assume the harmonics $w_{n\ne0}$ and $v_{n\ne0}$ to be of
181: order $\epsilon$. To lowest order ($\epsilon^1$), this gives the harmonics
182: \begin{eqnarray}
183:   w_{n\ne0} &=& (a_1g_nw_0 + dg_{n-1}w_0^*v_0)/(n\kappa), \nonumber\\
184:   v_{n\ne0} &=& (2a_2g_nv_0 + dg_{n+1}w_0^2)/(n\kappa).
185: \end{eqnarray}
186: Using these solutions, we obtain to first order $\epsilon$ the averaged 
187: equations for the DC components $w_0$ and $v_0$:
188: \begin{eqnarray}
189:   && i\frac{\partial w_0}{\partial z} +\frac{1}{2} 
190:      \frac{\partial^2 w_0}{\partial x^2} + \rho w_0^*v_0 +
191:      \gamma(|v_0|^2 - |w_0|^2)w_0 = 0, \nonumber\\
192:   && i\frac{\partial v_0}{\partial z} + \frac{1}{4} 
193:      \frac{\partial^2 v_0}{\partial x^2} + \tilde\beta v_0 + 
194:      \rho^* w_0^2 + 2\gamma|w_0|^2v_0 = 0.
195:   \label{average_eqns}
196: \end{eqnarray}
197: These equations also describe $m$th order QPM (where $\tilde\beta$=$\beta-m\kappa$
198: is ideally zero) and any other type of periodic grating, the parameters 
199: $\rho$ and $\gamma$ being simply given as sums over the Fourier coefficients 
200: of the grating \cite{ClaBanKiv97}. Incorporating time or the spatial $y$
201: coordinate is also straightforward.
202: For the square grating (\ref{FourGrating}),  $\rho$ and $\gamma$ can be explicitly 
203: calculated to
204: \begin{equation}
205:   \rho = i\frac{2d}{s\pi} + i\frac{4d_0(a_1 - a_2)}{s\pi\kappa}, \;  
206:   \gamma = \frac{d_0^2+d^2(1-8/\pi^2)}{\kappa}.  
207: \end{equation}
208: 
209: 
210: From Eqs.~(\ref{average_eqns}) follows the important result that {\em
211: cubic nonlinearities are induced in QNPCs} by nonlinear QPM gratings.
212: This cubic nonlinearity has the form of self- and cross-phase modulation 
213: (SPM and XPM), and is a result of non-phase-matched coupling between 
214: the wave at the main spatial frequency $\kappa$ and its higher harmonics. 
215: It is thus of a fundamentally different nature than the material 
216: Kerr nonlinearity, which is reflected in the fact that the SPM term 
217: is absent for the SH.
218: %The bright solitons of the average model (\ref{average_eqns}) have therefore 
219: %different properties than the bright solitons in $\chi^{(2)}$ media with a 
220: %competing focusing \cite{chi23_1D,Ban97} and defocusing cubic nonlinearity 
221: %\cite{BanBurKiv97}.
222: 
223: The averaged model (\ref{average_eqns}) is identical to the known model for 
224: nonlinear QPM gratings with no DC component ($a_j$=$d_0$=0), for which simulations 
225: confirmed that bright solitons had properties that were not predicted by the 
226: conventional model with only quadratic terms, but were accurately described 
227: with the inclusion of the cubic terms \cite{ClaBanKiv97}.
228: It was further shown that the induced cubic nonlinearity affects 
229: the phase modulation of cw waves, enabling efficient switching 
230: \cite{KobLedBanKiv98}, and that its strength can be increased
231: by modulating the grating \cite{BanClaChrTor99}. 
232: 
233: For the more general QNPCs considered here, the induced cubic nonlinearity 
234: depends on both the DC part and the modulation part of the nonlinear 
235: grating, but is independent of the linear grating. 
236: The cubic terms may lead to either a focusing or a defocusing effect, 
237: depending on the relative intensity of the fields and the sign of the 
238: phase mismatch $\beta$, since ${\rm sign}(\kappa)= {\rm sign}(\beta)$.  
239: The strength of the effective $\chi^{(2)}$ nonlinearity depends on the 
240: difference in the linear grating strengths at the fundamental and SH 
241: frequencies and on the DC component of the nonlinear grating.  
242: We thus recover the well-known effect that the interplay between the 
243: linear and nonlinear gratings can increase or decrease the effective 
244: $\chi^{(2)}$ nonlinearity, depending on the physical situation 
245: \cite{TanBey73,JasArvLau86}. 
246: 
247: The averaged model (\ref{average_eqns}) has stationary, localised soliton
248: solutions of the form $w_0(x,z)$=${\rm e}^{i \lambda z}\tilde w(x)/|\rho|$ 
249: and $v_0(x,z)$=${\rm e}^{2i\lambda z}\tilde v(x)/\rho$, which obey the 
250: equations
251: \begin{eqnarray}
252:   & & \frac{1}{2} \frac{\partial^2\tilde w}{\partial x^2} - \lambda\tilde w
253:       + \tilde w \tilde v + \tilde\gamma(\tilde v^2 - \tilde w^2)\tilde w 
254:       = 0, \nonumber\\
255:   & & \frac{1}{4}\frac{\partial^2 \tilde v}{\partial x^2} + 
256:       (\tilde\beta-2\lambda) \tilde v + \tilde w^2 + 
257:       2\tilde\gamma \tilde w^2\tilde v = 0,
258:   \label{soliton_eqns}
259: \end{eqnarray}
260: where $\tilde\gamma$=$\gamma/|\rho|^2$ depends only on $\kappa$, $a_1-a_2$,
261: and $d_0/d$.  
262: The slowly varying approximation gives valid solutions when the soliton period is longer than the grating period, i.e.~when the 
263: soliton parameter $\lambda$ is small, $|\lambda|\ll |\kappa|$.  
264: 
265: It is important to stress that Eqs.~(\ref{soliton_eqns}) cover a much more
266: general situation than in \cite{ClaBanKiv97}, which only considered a nonlinear grating.  
267: A given value of the parameter $\tilde\gamma$ represents {\em a range of 
268: physical situations with different combinations of linear and nonlinear 
269: gratings}.  
270: In Fig.~\ref{phys_grating} we illustrate representative combinations 
271: for exact phase matching ($\tilde\beta$=0) that all give the same 
272: value of $\tilde\gamma$.
273: 
274: \begin{figure}
275:   \centerline{\hbox{
276:   \psfig{figure=physgratings5.ps,width=\figwidthb,height=\figwidtha}}}
277:   \caption{QNPCs with the same value of $\tilde\gamma$. The linear 
278:   (nonlinear) grating is shown with a dashed (solid) curve.}
279:   \label{phys_grating}
280: \end{figure}
281: 
282: The first, simple case ($f$=1) is typical for domain inversion in 
283: ferroelectric materials, such as LiNbO$_3$. It has only a nonlinear 
284: grating that flips from positive to negative with no DC component, $d_0$=0.  
285: The second case has a DC level of $\chi^{(2)}$ nonlinearity, $d_0/d$=3,
286: but no linear grating, corresponding to the nonlinear part of the 
287: LiNbO$_3$/H:LiNbO$_3$ structure reported in \cite{Pet96}.  
288: The third case is the GaAs/GaAlAs structure reported in \cite{Pet96}, which
289: has a nonlinear grating with $d_0/d$=5/3 and a linear grating with 
290: $(a_1-a_2)/\beta$=$-0.07$.  
291: For the second and third cases to give the same $\tilde\gamma$ as the 
292: first, the grating wavenumber $\kappa$ must be multiplied by $f$=48.5 
293: and $f$=26.6, respectively. 
294: Physically this can be done by changing 
295: the input beam width $x_0$, thus maintaining $\tilde\beta$=0.
296: 
297: \begin{figure}
298:   \centerline{\hbox{
299:   \psfig{figure=nt_plots_a4.ps,width=\figwidtha}
300:   \psfig{figure=nt_plots_b4.ps,width=\figwidtha}}}
301:   \caption{Soliton properties versus the internal parameter $\lambda$ for 
302:   $\tilde\gamma$=0.02 (solid), $\tilde\gamma$=$-0.02$ (dashed), and 
303:   $\tilde\gamma$=0 (dotted), and three values of the residual mismatch 
304:   $\tilde\beta$. (a) Ratio of peak intensities $R$=$\tilde v^2(0)/\tilde 
305:   w^2(0)$. (b) Total power $P$.} 
306:   \label{sol_families}
307: \end{figure}
308: 
309: We have numerically found the bright soliton solutions of Eqs.~(\ref{soliton_eqns}) using a standard relaxation technique.
310: Bright soliton properties were investigated in \cite{ClaBanKiv97},
311: but not for normal dispersion, $\tilde\gamma>0$.
312: In Fig.~\ref{sol_families} we present the properties for normal 
313: ($\tilde\gamma$=0.02) and anomalous ($\tilde\gamma$=$-0.02$) dispersion, 
314: together with the zeroth-order solution ($\tilde\gamma$=0).
315: The ratio $R=\tilde v^2(0)/\tilde w^2(0)$ of peak intensities, shown 
316: in Fig.~\ref{sol_families}(a), confirms that the zeroth-order approximation 
317: becomes increasingly inaccurate for large $\lambda$.  
318: Also, for a given $\tilde\gamma$, $R$ approaches the same limiting value as 
319: $\lambda$ increases, regardless of the value of $\tilde\beta$.  
320: In this limit the SH is stronger than the fundamental for $\tilde\gamma>0$
321: ($R>1$) and much weaker for $\tilde\gamma<0$ ($R\simeq0$).  
322: The total power $P$=$\int_{-\infty}^\infty(\tilde v^2 + \tilde w^2)dx$ is 
323: shown in Fig.~\ref{sol_families}(b).  For $\beta > 0$ this reveals the interesting property that the
324: power threshold for existence, decreases for $\tilde\gamma>0$ 
325: and increases for $\tilde\gamma<0$, as compared to the zeroth-order value.
326:  
327: \begin{figure}
328:   \centerline{\hbox{\psfig{figure=cmds_plots_all.ps,width=\figwidthb}}}
329:   \caption{Ratio of peak intensities $R$=$\tilde v^2(0)/\tilde w^2(0)$ versus 
330:    $\lambda$ for $\kappa$=$10f$ (continuous), $\kappa$=$4f$ (dot-dashed), and 
331:    $\kappa$=$-10f$ (dashed), and for the zeroth-order solution (dotted). 
332:    The average of the propagating solitons are shown with $f$=1 (crosses), 
333:    $f$=48.5 (plusses), and $f$=26.6 (circles). $\tilde\beta$=0.} 
334:   \label{compare}
335: \end{figure}
336: 
337: The bright soliton solutions of the average model (\ref{average_eqns}) were 
338: tested for the three QNPCs of Fig.~\ref{phys_grating} by mapping them back 
339: to the variables $w$ and $v$, and using them as initial conditions in 
340: simulations of the field Eqs.~(\ref{field_eqns}).  
341: The evolution consists of small, regular oscillations superimposed on the slow 
342: average beam.  
343: Properties of the propagating solitons were calculated by averaging over
344: an integer number of grating periods and were then compared with the
345: predictions of the average model.  
346: Figure \ref{compare} displays the ratio of peak intensities versus $\lambda$
347: for exact phase matching, $\tilde\beta$=0, and reveals that, for both 
348: anomalous and normal dispersion, the solutions of the average model are
349: accurate for small $\lambda$ and large $|\kappa|$, as expected.  
350: Even when $|\kappa|$=4 the first-order solutions provide a much better 
351: fit than the zeroth-order solutions.  
352: Our analysis thus shows that bright solitons exist and propagate stably
353: in QNPCs with many types and combinations of linear and nonlinear QPM gratings.
354: 
355: However, this is provided one is careful and does not eliminate the effective 
356: $\chi^{(2)}$ nonlinearity by using a grating with $\kappa$=$2(a_2-a_1)d_0/d$. 
357: This could happen in realistic QNPCs without violating the assumption
358: $|\kappa|\gg1$.
359: For the LiNbO$_3$ and GaP/AIP structures given in \cite{Pet96} at 
360: phase matching ($\kappa$=$\beta$), the presence of the linear grating 
361: changes the effective $\chi^{(2)}$ nonlinearity by a factor of 
362: ${\cal F}$=$1+2d_0(a_1-a_2)/(d\kappa)$=1.4 and 0.3, respectively.  
363: The linear grating thus adds constructively in the LiNb0$_3$ structure and 
364: destructively in GaP/AIP.  In fact, modifying the nonlinear grating in the GaP/AIP structure slightly to 
365: $\chi^{(2)}_a$=40pm/V (max.) and $\chi^{(2)}_b$=19pm/V (min.) eliminates the effective 
366: $\chi^{(2)}$ nonlinearity entirely.
367: 
368: Conventional average models \cite{TanBey73,JasArvLau86,FejMagJunBye92} 
369: would predict that no soliton could exist with no nonlinearity, $\rho$=0.
370: However, in the model (\ref{average_eqns}) the induced cubic nonlinearity 
371: predicts that solitons should still exist as solutions of nonlinear 
372: Schr\"odinger equations.
373: In the case when the SH is 
374: strong ($v_0/w_0=\sqrt{5}\exp[i(\tilde\beta-\lambda/2)z]$), a family of bright solitons, $w_0$=$\sqrt{\lambda/(2\gamma)}\,{\rm sech}
375: (\sqrt{2\lambda}x){\rm e}^{i\lambda z}$,  exists for normal dispersion 
376: ($\gamma,\lambda$$>$0), and a family of dark solitons, $w_0=\sqrt{\lambda/
377: (4\gamma)}\,{\rm tanh}(\sqrt{|\lambda|}x){\rm e}^{i\lambda z}$, exists for 
378: anomalous dispersion ($\gamma,\lambda$$<$0).  With no SH, bright solitons, $w_0=\sqrt{2\lambda/|\gamma|}\,{\rm sech}
379: (\sqrt{2\lambda}x){\rm e}^{i\lambda z}$, exist for anomalous dispersion 
380: ($\gamma$$<$0,$\lambda$$>$0), whereas dark solitons, $w_0=\sqrt{|\lambda|
381: /\gamma}{\rm tanh}(\sqrt{|\lambda|}x) {\rm e}^{i\lambda z}$, exist for 
382: normal dispersion ($\gamma$$>$0,$\lambda$$<$0).
383: We test these solutions by mapping them back to the variables $w$ and $v$ 
384: and launching them as initial conditions in simulations of the field 
385: Eqs.~(\ref{field_eqns}).  
386: 
387: \begin{figure}
388:   \centerline{\hbox{
389:   \psfig{figure=bright5.ps,width=\figwidtha}
390:   \psfig{figure=dark5.ps,width=\figwidtha}}}
391:   \caption{(a) Bright and (b) dark solitons propagating in QNPCs with
392:   no effective $\chi^{(2)}$ nonlinearity. Shown is the scaled intensity of
393:   the fundamental for $|\kappa|$=100, $\tilde\beta$=0, and (a) $\gamma<0$,
394:   $\lambda$=1 and (b) $\gamma>0$, $\lambda$=$-1$. The SH is zero.} 
395: \label{darksoliton}
396: \end{figure}
397: 
398: Figure \ref{darksoliton} shows the evolution of a bright and dark soliton
399: with no SH over a distance of 25 grating periods.
400: The SH displays small, regular oscillations around the mean 
401: value zero, corresponding to the oscillations of the fundamental seen 
402: in Fig.~\ref{darksoliton}.
403: The simulations thus confirm that bright and dark solitons can indeed 
404: propagate in a stable manner in QNPCs with no effective $\chi^{(2)}$ 
405: nonlinearity. 
406: 
407: In summary, we have shown that bright solitons exist and propagate 
408: in a stable manner in 1D quadratic nonlinear photonic crystals (QNPCs) 
409: with many types and combinations of linear and nonlinear QPM gratings. 
410: %The gratings are QPM gratings, meaning that their band gap is well
411: %separated from the optical wavelength with Bragg reflections being 
412: %negligible.
413: By deriving first-order averaged equations, we have shown that such 
414: QNPCs have an induced cubic nonlinearity, and we have numerically found 
415: previously unknown families of bright solitons.
416: Even with no effective quadratic nonlinearity, the QNPCs support both 
417: bright and dark solitons due to the induced cubic nonlinearity. 
418: We have found analytical expressions for these solitons and shown that they
419: propagate in a stable manner.
420: 
421: Dark solitons are always unstable in homogeneous $\chi^{(2)}$ media in
422: settings for type I SHG, due to modulational instability of the back-ground 
423: plane-waves \cite{dark95}.
424: Our results show, for the first time, a dark soliton that appears to be stable
425: under propagation, with the stabilizing mechanism necessarily originating 
426: from the photonic crystal structure of the QNPC.
427: An analysis of dark solitons in the general case is being carried out.
428: 
429: The project is supported by the Danish Technical Research Council 
430: through Talent Grant No.~9800400.
431: 
432: \begin{thebibliography}{}
433: 
434: \bibitem{Sou96}
435:    NATO ASI Series: {\em Photonic Band Gap Materials}, ed.~C.M.~Soukoulis 
436:    (Kluwer Academic Publishers, 1996) 
437: 
438: \bibitem{Leo_etal00}
439: %  S.W. Leonard, J.P. Mondia, H.M. van Driel, O. Toader, S. John, K. Busch, 
440: %  A. Birner, U. G\"osele, and V. Lehmann,
441:    S.W. Leonard {\em et al.},
442: %  "Tunable two-dimensional photonic crystals using liquid-crystal 
443: %  infiltration",
444:    \prb {\bf 61}, R2389 (2000).
445: 
446: \bibitem{BusJoh99}
447:    K. Busch and S. John,
448: %  "Liquid-crystal photonic-band-gap materials: The tunable electromagnetic 
449: %  vacuum",
450:    \prl {\bf 83}, 967 (1999).
451: 
452: \bibitem{LarHibMizSte90}
453: %  S.~Larochelle, Y.~Hibino, V.~Mizrahi, G.I.~Stegeman,
454:    S.~Larochelle {\em et al}.,
455: %  "All-optical switching of grating transmission uing cross-phase
456: %  modulation in optical fibres",
457:    Electron.~Lett.~{\bf 26} (1990) 1459.
458: 
459: \bibitem{SteHagTor96}
460:    For an review see 
461:    G. Stegeman, D.J. Hagan, and L. Torner,
462: %  "$\chi^{(2)}$ cascading phenomena and their applications to all-optical 
463: %  signal processing, mode-locking, pulse compression and solitons",
464:    Opt. Quantum Electron. {\bf 28}, 1691 (1996).
465: 
466: \bibitem{ArmBloDucPer62}
467: %  J.A. Armstrong, N. Bloembergen, J. Ducuing and P.S. Pershan, 
468:    J.A. Armstrong {\em et al.}, 
469:    Phys. Rev. {\bf 127}, 1918 (1962).
470: 
471: \bibitem{TanBey73}
472:    C.L. Tang and P.P. Bey, 
473: %  "Phase matching in second-harmonic generation using artificial periodic
474: %  structures",
475:    {IEEE} J. Quantum Electron. {\bf 9}, 9 (1973).
476: 
477: \bibitem{JasArvLau86}
478:    B. Jaskorzynska, G. Arvidsson, and F. Laurell, 
479: %  "Periodic structures for phase-matching in second harmonic generation
480: %  in titanium lithium niobate waveguides",
481:    Spie: Integrated Optical Circuit Engineering {III} {\bf 251}, 221 (1986).
482: 
483: \bibitem{FejMagJunBye92}
484: %  M.M. Fejer, G.A. Magel, D.H. Jundt, and R.L. Byer, 
485:    M.M. Fejer {\em et al.}, 
486: %  "Quasi-phase-matched second harmonic generation: Tuning and tolerances",
487:    {IEEE} J. Quantum Electron. {\bf 28}, 2631 (1992).
488: 
489: \bibitem{NPC_chi2_bragg}
490: %  A.~Ashkin, A.~Yariv,
491: %  "Bell Labs.~Tech.~Memo~No.~MM-61-124-46", Nov.~13, 1961.
492:    N.~Bloembergen and A.J.~Sievers,
493: %  "Nonlinear properties of periodic laminar structures",
494:    Appl.~Phys.~Lett.~{\bf 17} (1970) 483.
495: %  J.P.~van der Ziel and M.~Ilegems,
496: %  "Optical second harmonic generation in periodic multilayer
497: %  GaAs-Al$_{0,3}$Ga$_{0,7}$As structures",
498: %  Appl.~Phys.~Lett.~{\bf 28} (1976) 437.
499: 
500: \bibitem{Tor98}
501:    For a review see 
502:    L. Torner, in {\em Beam Shaping and Control with Nonlinear Optics}, 
503:    eds. F. Kajzer and R. Reinisch (Plenum, New York, 1998).
504: 
505: \bibitem{gap}
506:    Yu.S. Kivshar, 
507: %  "Gap solitons due to cascading",
508:    \pre {\bf 51}, 1613 (1995). 
509:    For a recent overview see
510:    A. Arraf and C.M. de Sterke, 
511: %  "Coupled-mode equations for quadratically nonlinear deep gratings",
512:    \pre {\bf 58}, 7951 (1998).
513: %  T. Peschel, U. Peschel, F. Lederer, and B. Malomed, \pre {\bf 55},
514: %  4730 (1997).
515: %  C. Conti, S. Trillo, and G. Assanto, \prl {\bf 78}, 2341 (1997).
516: %  H. He and P.D. Drummond, \prl {\bf 78}, 4311 (1997).
517: 
518: \bibitem{Hel00}
519: %  D.C. Hutchings, \apl {\bf 76}, 1362 (2000); 
520: %  D.C. Hutchings, private communication.
521: %  A. Saher Helmy, D.C. Hutchings, T.C. Kleckner, J.H. Marsh, A.C. Bryce,
522: %  J.M. Arnold, C.R. Stanley, J.S. Aitchison, C.T.A. Brown, K. Moutzouris,
523: %  and M. Ebrahimzadeh,
524:    A.A. Helmy {\em et al.},
525:    "Quasi-phase-matching in GAAS-ALAS superlattice waveguides via bandgap
526:    tuning using quantum well intermixing",
527:    \ol, {\em to appear}.
528: 
529: \bibitem{ClaBanKiv97}
530:    C.B. Clausen, O. Bang and Y.S. Kivshar, \prl {\bf 78}, 4749 (1997). 
531: 
532: \bibitem{DMsol}
533: %  For a review see S.K. Turitsyn, V.K. Mezentsev, and E.G. Shapiro,
534: %  "Dispersion-managed solitons and optimization of the dispersion management",
535: %  Opt. Fiber Tech. {\bf 4}, 384 (1998).
536: %  J.H.B. Nijhof, N.J. Doran, W. Forysiak, and F.M. Knox,
537:    J.H.B. Nijhof {\em et al.},
538: %  "Stable soliton-like propagation in dispersion managed systems with net
539: %  anomalous, zero and normal dispersion",
540:    Electron. Lett. {\bf 33}, 1726 (1997).
541:    S. K. Turitsyn and E. G. Shapiro, 
542: %  "Dispersion-managed solitons in optical amplifier transmission systems
543: %  with zero average dispersion", 
544:    \ol {\bf 23}, 682 (1998)
545: 
546: \bibitem{MenSchTor94}
547:    C.R. Menyuk, R. Schiek, and L. Torner, \josab {\bf 11}, 2434 (1994).
548:    O. Bang, {\em ibid.} {\bf 14}, 51 (1997).
549: 
550: \bibitem{KobLedBanKiv98}
551: %  A. Kobyakov, F. Lederer, O. Bang, and Y.S. Kivshar,
552:    A. Kobyakov {\em et al.},
553:    \ol {\bf 23} (1998) 506.
554: 
555: \bibitem{BanClaChrTor99}
556: %  O. Bang, C.B. Clausen, P.L. Christiansen, and L. Torner, 
557:    O. Bang {\em et al.}, 
558:    \ol {\bf 24}, 1413 (1999).
559: 
560: 
561: \bibitem{Pet96}
562:    D.V. Petrov, 
563: %  "Nonlinear phase shift by cascaded quadi-phase-matched second harmonic
564: %  generation",
565:    \oc {\bf 131}, 102 (1996).
566: 
567: \bibitem{dark95}
568:    S. Trillo and P. Ferro, \ol {\bf 20}, 438 (1995).
569:    A.V. Buryak and Yu.S. Kivshar, \ol {\bf 20}, 834 (1995).
570: 
571: %\bibitem{chi23_1D}
572: %   A.V. Buryak, Yu.S. Kivshar, and S. Trillo, \ol {\bf 20}, 1961 (1995); 
573: %   S. Trillo, A.V. Buryak, and Yu.S. Kivshar, \oc {\bf 122}, 200 (1996);
574: %   M.A. Karpierz, \ol {\bf 20}, 1677 (1995).
575: 
576: %\bibitem{BanBurKiv97}
577: %   O. Bang, Yu.S. Kivshar, A.V. Buryak, \ol {\bf 22}, 1680 (1997).
578: \end{thebibliography}
579: 
580: \end{multicols}
581: 
582: \end{document}
583: