1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % File: OLpagelength.tex %
3: % Date: Feb. 14, 2007 %
4: % %
5: % For assistance, contact Joseph Richardson, %
6: % jricha@osa.org %
7: % %
8: % LaTeX template and instructions for %
9: % length check and submission of OSA %
10: % Optics Letters manuscripts %
11: % %
12: % %
13: % \documentclass[10pt,letterpaper,twocolumn]{article} %
14: % \usepackage{ol2} %
15: % %
16: % (c) 2007 Optical Society of America %
17: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
18:
19: %\documentclass[10pt,letterpaper,twocolumn]{article} %% two column, final layout
20:
21: \documentclass[12pt]{article} % single column, double spaced
22: \usepackage[tablesfirst,notablist,nomarkers]{endfloat} %% float figs. to back
23:
24: \usepackage{ol2}
25: \usepackage[draft]{hyperref}
26: \usepackage{amsmath}
27:
28:
29: \begin{document}
30:
31: %\twocolumn[ %% activate for two-column option
32:
33: \title{Tunable photonic Bloch oscillations in electrically modulated photonic crystals}
34:
35: %% For REVTeX it is possible to automate superscript and e-mail callouts with the superscriptaddress option; see REVTeX4 documentation.
36:
37:
38: \author{Gang Wang,$^{1,2}$ Ji Ping Huang,$^{1,2,*}$ and Kin Wah Yu$^{1,**}$}
39:
40: \address{
41: $^1$Department of Physics, Chinese
42: University of Hong Kong, Shatin, New Territories, Hong Kong
43: \\
44: $^2$Surface Physics Laboratory and Department of Physics, Fudan
45: University, Shanghai 200433, China \\
46: $^*$Corresponding author: jphuang@fudan.edu.cn\\
47: $^{**}$Corresponding anthor: kwyu@phy.cuhk.edu.hk
48: }
49:
50: \begin{abstract}
51:
52:
53: We exploit theoretically the occurrence and tunability of photonic
54: Bloch oscillations (PBOs) in one-dimensional photonic crystals
55: (PCs) containing nonlinear composites. Because of the enhanced
56: third-order nonlinearity (Kerr type nonlinearity) of composites,
57: photons undergo oscillations inside tilted photonic bands, which
58: are achieved by the application of graded external pump electric
59: fields on such PCs, varying along the direction perpendicular to
60: the surface of layers. The tunability of PBOs (including amplitude
61: and period) is readily achieved by changing the field gradient.
62: With an appropriate graded pump AC or DC electric field, terahertz
63: PBOs can appear and cover a terahertz band in electromagnetic
64: spectrum.
65:
66: \end{abstract}
67:
68: \ocis{230.4910, 350.4238, 350.5500.}
69:
70: %] %% activate for two-column option
71:
72: \noindent Recently matter wave dynamics of Bloch oscillations has
73: already motivated a good deal of attention. In 1928, Bloch made a
74: striking prediction that in a crystal lattice, a homogeneous static
75: electric field induces an oscillatory rather than uniform motion of
76: the electrons~\cite{Bloch}, referred to as Bloch oscillations (BOs).
77: Electronic BOs have been observed in semiconductor
78: superlattices~\cite{Esaki,Feldmann}. Now this phenomenon has been
79: extended to various classical wave systems, such as acoustic
80: systems~\cite{AlepuzPRL2007}, elastic
81: systems~\cite{MoralesPRL2006}, and photonic
82: systems~\cite{SapienzaPRL2003}. The
83: electronic BOs are quite easily tunable via externally applied
84: electric fields or magnetic fields, or simultaneously by
85: both~\cite{Grill2007}. However, for uncharged particles like
86: photons, one has to resort to other approaches for
87: controllability~\cite{FanPRB2005}.
88:
89:
90: In this work, we propose a class of PCs to realize the occurrence
91: and tunability of photonic Bloch oscillations (PBOs) by graded pump
92: DC or AC electric fields ${\bf E}_0$. The idea relies on the fact
93: that the effective dielectric permittivity (thus refractive index)
94: of some materials depends on external electric fields because of the
95: third-order nonlinearity (Kerr type nonlinearity). To exemplify this idea,
96: we theoretically calculate the PBOs in such PCs under different
97: graded pump electric fields. Compared with existing
98: proposals~\cite{FanPRB2005}, our proposal offers ultrafast response
99: time and dynamical control. Also, we show that with an
100: appropriate choice of parameters, this could lead to a generation of
101: teraherz radiations.
102:
103: Figure 1 shows a one-dimensional double-layer PC, which consists of
104: alternative layers of composite materials~\cite{Wang} (or composites
105: for short, e.g., Au/SiO2) and common dielectric (e.g., air). Through
106: local field and resonant scattering effects in
107: nanoparticles~\cite{HuangPhysRep2006}, the third-order
108: nonlinearity of composites can be
109: enhanced~\cite{RichardOptlett1985,HuangPhysRep2006}.
110:
111:
112: Let us denote by ${\bf D}_{0 (1)}$ the response to pump (probe)
113: field ${\bf E}_{0 (1)}$. For weak nonlinearity under consideration,
114: namely $\chi^{(3)}{\bf |E|}^2\ll\epsilon$, the constitutive relation
115: ${\bf D}=\epsilon {\bf E} + \chi^{(3)} {\bf |E|}^2 {\bf E}$ can be
116: written as ${\bf D}={\bf D}_0+\alpha {\bf D}_1$, and ${\bf E}={\bf
117: E}_0+\alpha {\bf E}_1$, where $\alpha$ is a small parameter. Due to
118: the weak nonlinearity limit, we apply the perturbation
119: method~\cite{YuPRB1992} to express ${\bf D}_0$ and ${\bf D}_1$ in
120: terms of ${\bf E}_0$ and ${\bf E}_1$: ${\bf
121: D}_0=(\epsilon+\chi^{(3)} |{\bf E}_0|^2){\bf E}_0$, ${\bf
122: D}_1=\epsilon {\bf E}_1+\chi^{(3)} |{\bf E}_0|^2 {\bf
123: E}_1+2\chi^{(3)} \rm{Re[E_0^*E_1]}\cos\theta{\bf E}_0$. Since the
124: pump field ${\bf E}_0$ is much larger than the probe field ${\bf
125: E}_1$, the response to the probe field ${\bf D}_1$ is rather
126: stable, and it is related to the angle $\theta$ between the pump
127: field (${\bf E}_0$) and the probe field (${\bf E}_1$). Without loss
128: of generality, we assume ${\bf E}_1\| {\bf E}_0$ (see
129: Fig.~\ref{fig1}), which results in an effective dielectric constant
130: for the probe field
131: $\tilde{\epsilon}_{eff}=\epsilon+3\chi^{(3)}|{\bf E}_0|^2$. So the
132: dielectric constant $\tilde{\epsilon}_p$ possessed by nonlinear
133: nanoparticles can be expressed as
134: \begin{equation}
135: \tilde{\epsilon}_p=\epsilon_{p}+3\chi_p^{(3)}|\textbf{E}_p|^2
136: \approx
137: \epsilon_{p}+3\chi_p^{(3)}\langle|\textbf{E}_p|^2\rangle,\label{eq1}
138: \end{equation}
139: where $\epsilon_{p}$ denotes the linear (field-independent)
140: dielectric permittivity, $\chi_p^{(3)}$ the third-order nonlinear
141: susceptibility of the nanoparticles, ${\bf E}_p$ the local electric
142: field inside the nanoparticle, and $\langle\cdots\rangle$ the volume
143: average of $\cdots$. Then, the effective nonlinear permittivity
144: $\tilde{\epsilon}_{1}$ of the layer of composites can be given by
145: the Maxwell-Garnett
146: approximation~\cite{HuangPhysRep2006}
147: \begin{equation}
148: \frac{\tilde{\epsilon}_{1}-\epsilon_{h}}{\tilde{\epsilon}_{1}+2\epsilon_{h}}=p
149: \frac{\tilde{\epsilon}_{p}-\epsilon_{h}}{\tilde{\epsilon}_{p}+2\epsilon_{h}},
150: \end{equation}
151: where $\epsilon_{h}$ represents the dielectric permittivity of the
152: host (which is assumed to be linear for simplicity), and $p$ the
153: volume fraction of nanoparticles. The volume-averaged local electric
154: field $\langle|{\bf E}_p|^2\rangle$ in the layer is given
155: by~\cite{ShengPingJOSAB1998,YuSSC98}
156: \begin{equation}
157: \langle|{\bf E}_p|^2\rangle =
158: \frac{9|\epsilon_{h}|^2}{|(1-p)\epsilon_p+(2+p)\epsilon_{h}|^2}{\bf
159: E}_0^2\equiv \beta^2 {\bf E}_0^2.\label{equation}
160: \end{equation}
161: Thus the effective nonlinear response of the layer can be much
162: enhanced.
163:
164: For model calculations, we investigate the composite of Au/SiO$_2$
165: with volume fraction of Au nanoparticles $p=0.20$. The dielectric
166: permittivities of SiO$_2$ and air (dielectric layer) are taken to be
167: $\epsilon_h = 2.25\epsilon_0$ and $\epsilon_{air} = \epsilon_0$. The
168: probe field can be much weaker in strength than the pump electric
169: field, so that the dispersion (and loss) of the (much weaker) probe
170: field can be neglected. Here we take
171: $\epsilon_p=-9.97\epsilon_0$~\cite{ShengPingAPL1997,ShengPingJOSAB1998},
172: which is a real frequency-independent constant. This corresponds to
173: a pump field produced by a laser source at
174: $620\,$nm~\cite{ShengPingJOSAB1998}.
175:
176:
177: The existence of the enhanced Kerr nonlinearity enables us to
178: modify photonic band structures of such PCs by applying pump
179: electric fields~\cite{Wang}. Throughout this work, we use
180: $\chi_p^{(3)}E_0^2$ to indicate the strength of external pump
181: electric field $E_0$. To obtain PBOs, we apply an external electric
182: field with gradation profile $\chi_p^{(3)}E_0^2=gz+b$, which can be
183: adjusted to achieve different gradation profiles. In graded fields,
184: the photonic band structures become depth-dependent, viz.,
185: $z$-dependent (see Fig.~\ref{fig2}), which can be evaluated via the
186: transfer matrix method with $g=3.67\times10^{-4}$ and
187: $b=1.33\times10^{-2}$. The reason is easily understood. When the
188: pump electric field acting on the layers of composites varies along
189: the $z$ direction, the effective dielectric permittivity of each
190: composite layer grows monotonically, resulting in a refractive index
191: gradient. By carefully examining the condition for the appearance of
192: PBOs, we are convinced that PBOs cannot happen in the lowest band.
193: Thus we will study PBOs in the second (or higher) band. Besides the
194: tilted band structures in Fig.~\ref{fig2}, the other prerequisite
195: for the appearance of PBOs is that $\omega(z_{max},~k=0)
196: =\omega(z_{min},~k=\pi/a)$, where $\omega$ represents the angular
197: frequency of source waves, $z_{max(min)}$ means the maximum
198: (minimum) position for PBOs, and $k$ is the Bloch wave vector. In
199: other words, PBOs just appear within the cross section that is
200: composed by the two horizontal dotted lines as displayed in
201: Fig.~\ref{fig2}. So when the incidence with $\omega_0=1.76$ $c/a$
202: illuminates the structure (see the arrow in Fig~\ref{fig2}), the
203: oscillations occure in the spatial range $250a< z< 500a$, in good
204: agreement with the numerical results shown in Fig.~\ref{fig4} (a).
205:
206:
207: The above analysis shows that PBOs can appear inside the second band
208: if a graded external electric field is applied on such 1D PCs. When
209: electromagnetic waves are incident on the PCs, multiple reflections
210: on the gaps lead to spatial BOs. The actual calculation is to solve
211: for the propagation of a Gaussian wavepacket initially peaked around
212: $z=z_{max}$ and $k=0$ at $t=0$. The subsequent motion of the
213: wavepacket is governed by the Hamiltonian equations of motion of
214: $\omega(z,k)$. At $t=T_B/2$ ($T_B$: time period of PBOs), $z$
215: becomes $z_{min}$ while $k$ reaches $\pi/a$ and PBO occurs. The
216: Hamiltonian equations of motion of $\omega(z,k)$ have been
217: integrated with appropriate initial conditions to yield the time
218: series of displacement $z(t)$ and momentum $k(t)$ (not shown
219: herein). As expected, such oscillations are clearly seen from the
220: mean position $\langle z(t) \rangle$ in Fig.~\ref{fig4} (a) and (b).
221: For the use and justification of the Hamiltonian formalism for
222: optics, please see Ref.~\cite{Russell} and references therein. Based
223: on the semiclassical solutions, we access the width $\Delta z(t)$ of
224: a wave packet and show that it remains bound in time. For a packet
225: with Gaussian distribution with widths $\sigma_z = 5$ and $\sigma_k
226: = 0.2$, we show the time dependent mean width $\Delta z(t)$
227: Fig.~\ref{fig4} (c) and (d). We can find that the width is about
228: 5$\%$ of the length of the superlattice (600$a$) and increases only
229: slightly after several periods, which is acceptable. This originates
230: from the non-constant inclination of the band diagram. Thus we can
231: be convinced that Bloch oscillation indeed occurs. In fact, we can
232: design linearly tilted bands to avoid the increase of the width.
233:
234: The enhancement of the third-order nonlinearity of the composites
235: enables us to tune PBOs by using tunable inclined band structures.
236: Figure~\ref{fig4} shows the spatial range of PBOs for a fixed
237: incident frequency $\omega_0=1.76$ $c/a$ under different gradation
238: profiles. We find that the variation of gradient plays an important
239: role in the occurrence of PBOs, including their amplitude and
240: oscillation frequency. A key parameter for PBOs is oscillation
241: period $T_B$. Figure~\ref{fig4} also shows that $T_B$ depends on the
242: tuning parameters in the gradation profiles. Apparently, $T_B$
243: decreases while increasing the field gradient because of steeper
244: tilting bands. Therefore, there is a critical gradient
245: $g=2.73\times10^{-4}$, below which $T_B$ becomes infinite (or,
246: alternatively, no oscillations come to appear).
247:
248: Considering the large nonlinear susceptibility (typical value $8
249: \times 10^{-8}$ esu) of Au nanoparticles~\cite{ShengPingJOSAB1998},
250: for the two gradation profiles adopted in our manuscript
251: $|E|^2=(gz+b)/\chi^{(3)}$, the maximum pump fields intensities are
252: $I_m=|E|^2= 61.5$ and $72.1$ mW/cm$^2$ respectively. They are
253: experimentally practical. The intensity of probe fields
254: can be fraction of milliwatts per cm$^2$, which is the requirement
255: of the weak nonlinearity. When the probe fields are at 535 nm, and
256: the thickness of each layer of the superlattice $d_1=d_2=75\,$nm,
257: the BO frequencies $f_B = 1/T_B$ are 0.720\,THz and 0.853\,THz
258: respectively. All the $f_B$'s are just located within the range of
259: terahertz band, namely, $10^{11} -10^{13}\,$Hz. For decoupling THz
260: radiation from the structure, one can apply a uniform pump field to
261: the structure, leading to the non-tilted band diagram. Meanwhile,
262: the structure is illuminated by an incident light with frequency in
263: the region of photonic allowed bands. Second, once the graded pump
264: electric field is applied suddenly, the pulse propagating in the
265: structure will be trapped and the oscillations commence
266: subsequently. Third, after several oscillation periods, the graded
267: pump field is restored to the uniform case, then the carrier waves
268: with terahertz modulation can escape from the structure. Meanwhile,
269: to avoid the leakage of energy in the transverse direction, spatial
270: confinement is needed. In this regard, total internal reflection
271: like in graded-index optical fibers can be used. For low-index
272: medium, it is useful to guide light by means of a photonic band gap.
273:
274: %Here some comments are in order. In fact, the tunability of PBOs can
275: %also be achieved by choosing different incident frequencies because
276: %of the nonlinearly inclined depth-dependent band structures. Since
277: %the slope of the tilted band is not a constant, different incident
278: %beam frequencies cause the shift of spatial oscillation centers,
279: %which leads to various $f_B$. It is worth noting that this is a
280: %classical wave system where the incident frequency can be continuous
281: %provided that it satisfies the condition for PBOs to occur. This is
282: %unlike quantum systems which possess discrete energy levels. On the
283: %other hand, the transmission through the crystal is determined by
284: %resonant Zener tunneling. This is quite reminiscent of quantum
285: %systems. The resonant Zener tunneling provides another means to
286: %extract PBOs from the crystal as well~\cite{FanPRB2005}.
287:
288: In summary, we have theoretically exploited the occurrence and
289: tunability of PBOs in
290: one-dimensional PCs containing nonlinear
291: composites by applying graded external electric fields.
292: Meanwhile, a kind of terahertz PBOs can appear and cover a
293: terahertz band in electromagnetic spectrum.
294:
295: {\it Acknowledgments.} This work was supported by the RGC Earmarked
296: Grant of Hong Kong SAR Government, by the C. N. Yang Fellowship in
297: CUHK, by the CNKBRSF under Grant No. 2006CB921706, by the Pujiang
298: Talent Project (No. 06PJ14006) of the Shanghai Science and
299: Technology Committee, by the Shanghai Education Committee ("Shu
300: Guang" project), and by the NNSFC under Grant No. 10604014.
301:
302:
303:
304: \begin{figure}[htb]
305: \centerline{\includegraphics[width=8cm]{PC.eps}} \caption{(Color
306: online) Schematic view of the one-dimensional PC composed of a
307: composite layer (nonlinear composites) with thickness $d_1$ and a
308: dielectric layer (e.g., air) with thickness $d_2$. The composites
309: can be prepared with nonlinear nanoparticles (e.g., gold)
310: randomly dispersed in a linear dielectric (e.g., silica), as
311: shown in the right panel. For plotting
312: Figs.~\ref{fig2}-\ref{fig4}, we shall use $d_1=d_2=0.5a$, where
313: $a$ denotes the lattice constant.}\label{fig1}
314: \end{figure}
315:
316:
317: \begin{figure}[htb]
318: \centerline{\includegraphics[width=6cm]{energy.eps}} \caption{(Color
319: online) Depth-dependent photonic band structure for gradation
320: profile $\chi^{(3)}_p E_0^2=g z+b$ with parameters
321: $g=3.67\times10^{-4}$ and $b=1.33\times10^{-2}$ [which correspond to
322: Fig.~\ref{fig4}(a)]. The vertical dashed lines label the space of
323: oscillations displayed in Fig.~\ref{fig4}(a), for an incident wave
324: with $\omega_0=1.76$ $c/a$. The cross section composed by the two
325: horizontal dotted lines in the second band denotes the full region
326: for the appearance of oscillations for various angular frequencies
327: of incident electromagnetic waves.}\label{fig2}
328: \end{figure}
329:
330:
331: \begin{figure}[htb]
332: \centerline{\includegraphics[width=8cm]{x.eps}} \caption{(Color
333: online) The time-dependent mean position and width of a wave
334: packet under pump electric fields with different gradation
335: profiles $\chi^{(3)}_p E_0^2=gx+b$. The panels (a) and (c)
336: correspond to $g=3.67\times10^{-4}$ and $b=1.33\times10^{-2}$,
337: (b) and (d) for $g=4.33\times10^{-4}$ and $b=1.33\times10^{-2}$.
338: The Bloch oscillations clearly develop around different centers
339: with different periods.}\label{fig4}
340: \end{figure}
341:
342:
343:
344:
345: \begin{thebibliography}{99}
346:
347:
348: \bibitem{Bloch} F. Bloch, Z. Phys. {\bf 52}, 555 (1928); C. Zener, Proc. R. Soc. London A {\bf 145}, 523 (1934).
349:
350: \bibitem{Esaki} L. Esaki and R. Tsu, IBM J. Res. Dev. {\bf 14}, 61 (1970).
351:
352: \bibitem{Feldmann} C. Waschke, H. G. Roskos, R. Schwedler, K. Leo, H. Kurz, and K. K\"ohler, Phys.
353: Rev. Lett. {\bf 70}, 3319 (1993).
354:
355: \bibitem{AlepuzPRL2007} H. Sanchis-Alepuz, Y. A. Kosevich, and J. S\'{a}chez-Dehesa, Phys. Rev. Lett. {\bf 98}, 134301 (2007).
356:
357: \bibitem{MoralesPRL2006} L. Guti\'{e}rrez, A. D\'{i}az-de-Anda, J. Flores, R. A.
358: M\'{e}ndez-S\'{a}nchez, G. Monsivais, and A. Morales, Phys. Rev.
359: Lett. {\bf 97}, 114301 (2006).
360:
361: \bibitem{SapienzaPRL2003} R. Sapienza, P. Costantino, D. Wiersma, M. Ghulinyan, C. J. Oton, and L. Pavesi, Phys. Rev. Lett. {\bf 91}, 263902 (2003).
362:
363: \bibitem{Grill2007} L. Smr\v{c}ka, N. A. Goncharuk, M. Orlita, and R. Grill, Phys. Rev. B {\bf 76}, 075321 (2007).
364:
365: \bibitem{FanPRB2005} V. Lousse and S. Fan, Phys. Rev. B {\bf 72}, 075119 (2005).
366:
367: \bibitem{Wang} G. Wang, J. P. Huang, and K. W. Yu, Appl. Phys. Lett. {\bf 91}, 191117 (2007).
368:
369: \bibitem{HuangPhysRep2006} J. P. Huang and K. W. Yu, Phys. Rep. {\bf 431}, 87 (2006).
370:
371: \bibitem{RichardOptlett1985} D. Ricard, P. Roussignol, and C. Flytzanis, Opt. Lett. {\bf 10}, 511 (1985).
372:
373: \bibitem{YuPRB1992} G. Q. Gu and K. W. Yu, Phys. Rev. B {\bf 46}, 4502 (1992).
374:
375: \bibitem{ShengPingJOSAB1998} H. R. Ma, R. F. Xiao, and P. Sheng, J. Opt. Soc. Am. B {\bf 15}, 1022 (1998).
376:
377: \bibitem{YuSSC98} K. W. Yu, Solid State Commun. {\bf 105}, 689 (1998).
378:
379: \bibitem{ShengPingAPL1997} H. B. Liao, R. F. Xiao, J. S. Fu, P. Yu, G. K. L. Wong, and P. Sheng, Appl. Phys. Lett. {\bf 70}, 1 (1997).
380:
381: \bibitem{Russell} P. St. J. Russell and T. A. Birks, J. Lightwave Technol. {\bf 17}, 1982 (1999).
382:
383: \end{thebibliography}
384:
385:
386: \end{document}
387: