Brillouin-zone spectroscopy of nonlinear photonic

lattices

Guy Bartal,1 Oren Cohen,1 Hrvoje Buljan,1,2 Jason W. Fleischer,1,3 Ofer Manela,1

Mordechai Segev1

1Physics Department, Technion - Israel Institute of Technology, Haifa 32000, Israel

2Department of Physics, University of Zagreb, PP 322, 10000 Zagreb, Croatia
3Electrical Engineering Department, Princeton University, New Jersey 08544

Abstract

We present a novel, real-time, experimental technique for linear and nonlinear Brillouin

zone spectroscopy of photonic lattices. The method relies on excitation with random-

phase waves and far-field visualization of the spatial spectrum of the light exiting the

lattice. Our technique facilitates mapping the borders of the extended Brillouin zones and

the areas of normal and anomalous dispersion within each zone. For photonic lattices

with defects (e.g., photonic crystal fibers), our technique enables far-field visualization of

the defect mode overlaid on the extended Brillouin zone structure of the lattice. The

technique is general and can be used for photonic crystal fibers as well as for periodic

structures in areas beyond optics.

 It is very often that the characteristics of waves propagating in periodic structures are

best described in Fourier space. While waves in periodic structures occur throughout

science, photonic systems have the advantage that optical inputs can be easily

manipulated and results can be directly imaged. Moreover, optics provides many

technologically important systems relying on the dynamics in periodic structures, such as

photonic crystals [1,2,3], photonic crystal fibers (PCF) [4,5,6] and waveguide arrays. The

periodicity partitions the linear transmission spectrum of waves propagating in such

structures into bands of propagating (Bloch) modes separated by forbidden gaps. The

lattice geometry, and the corresponding wave dynamics, is most efficiently presented in

the Fourier (momentum) space through the extended Brillouin zone map [7], whose

boundaries are defined by the Bragg-reflection planes. For example, Fig. 1a shows the

first and second Brillouin zones of a square lattice with the high-symmetry points (Γ, X,

and M) marked by white dots. Within each band, the curvature of the transmission

spectrum varies with the spatial frequency (transverse momentum) and changes sign.

This results in regions of normal and anomalous dispersion/diffraction in exactly the

same way as band curvature causes positive and negative effective mass in solid-state

physics. The first two bands in the transmission spectrum are shown in Fig. 1b, with the

dispersion curves between the symmetry points of the first two bands plotted in Fig. 1c.

The equivalence between optical waves in a periodic dielectric structure and electrons

in a periodic atomic potential has opened up the areas of "Bloch wave optics" [8] and

“photonic bandgap engineering” [2]. For example, photonic crystals can exhibit a gap in

the electromagnetic spectrum: a range of frequencies at which light cannot propagate. In a

similar vein, Bragg diffractions from such structures give rise to omni-directional mirrors

reflecting all incident waves within a wide angular range, manifesting a gap in the spatial

transmission spectrum [9]. These ideas, combined with a dielectric defect embedded in

the periodic structures, have allowed the fabrication of photonic crystal fibers (PCFs)

[10] and photonic bandgap lasers [11]. In nonlinear photonic lattices, the defect can be

self-induced, leading to the formation of lattice solitons [12, 13, 14]. Nonlinearity also

serves to couple linear modes, making nonlinear photonic systems excellent platforms to

study (image) fundamental problems in lattice dynamics.

Here, we present a novel experimental technique for linear and nonlinear Brillouin

zone spectroscopy of photonic lattices. Our technique facilitates a direct visualization of

the lattice structure in Fourier space by mapping the borders of the extended Brillouin

zones, and marking the areas of normal and anomalous dispersion within them. The

method relies on excitation of the lattice modes with partially-incoherent waves,

performing an optical Fourier transform and measuring the power spectrum. The

technique is general and can be used for any periodic structure. We experimentally

demonstrate it on square and hexagonal waveguide arrays, and also study 2D waveguide

arrays with positive and negative defects embedded in their structure (equivalent to solid-

core and hollow-core photonic crystal fibers, respectively). The method provides a

powerful diagnostic tool for photonic lattices. Equally important, our technique points the

way towards future experiments on nonlinear mode coupling which have been intriguing

scientists since the pioneering work of Fermi, Pasta, and Ulam in 1955 [15].
 To map out the momentum space of a lattice, the lattice should be probed by a broad

spectrum of its eigen-modes (Bloch waves) in the mapped region of k-space. Preferably,

one would like to excite the lattice with a broad and uniform spectrum of Bloch modes,

residing in several different bands. At the same time, the probe beam must also be broad

enough in real space to sample a large enough number of unit cells. These requirements

can be reconciled simultaneously by a random-phase (spatially incoherent) probe beam,

which facilitates homogeneous excitation of several BZs with a beam that occupies

numerous lattice sites. Moreover, the time-averaged intensity of such an incoherent probe

beam, which excites many Bloch modes, is smooth, because the speckles are washed out.

For these reasons, we probe the photonic lattices with a partially-spatially-incoherent

(random-phase) beam, which has a uniform spatial power spectrum extending over

several Brillouin zones, and is broad enough to cover numerous lattice sites.

Our experiments are carried out in optically-induced photonic lattices [13], formed in

a highly anisotropic photosensitive nonlinear crystal, utilizing the methods that have

recently proven to be very successful for spatial solitons experiments in photonic lattices

[13,16]. The experimental setup is shown in Fig. 1d. In all experiments, we use a 488 nm

laser beam. The 2D square lattice is induced by interfering two pairs of plane waves in

the nonlinear material, while the hexagonal (trigonal) lattice is induced by interfering

three plane waves under similar conditions [Fig 2]. In both cases, the lattice-forming

waves are polarized so as to propagate linearly in the nonlinear medium, serving only to

induce the periodic photonic structure. Our partially-spatially-incoherent probe beam is

generated by passing a laser beam through a rotating diffuser, and the power spectrum of
the emerging beam is controlled using a filter in the focal plane of a lens. The laser beam

exiting the diffuser is imaged onto the input face of the nonlinear crystal in which the

photonic lattice is induced. The data is taken by monitoring (photographing) the intensity

distribution at the focal plane of the “exit lens” positioned so as to visualize the Fourier

power spectrum of the light exiting the crystal.

Consider first the linear scheme for mapping the extended BZ of square and hexagonal

lattices, both with ~10 µm period. The probe beam is 25 µm FWHM, having a transverse

correlation distance of 5 µm . Typical results are shown in Fig. 2, where the upper (lower)

row depicts experimental data obtained with a square (hexagonal) lattice. The left column

(a,e) shows the interference pattern of the array waves forming the lattice, as they exit the

nonlinear crystal. The second column (b,f) shows the power spectra of the input

incoherent probe beam (wide homogeneous circle of illumination) illuminating the

lattice, with the far-field of the array-forming waves (the sharp peaks) marking the

corners of the respective first Brillouin zones (the incoming plane waves define the

corresponding Bragg angles). The third row (c,g) shows the text-book calculated picture

of the extended Brillouin zones of the respective lattices. And finally, the right column

(d,h) shows the experimental far-field picture of the incoherent light emerging from the

lattice, depicting the first four Brillouin zones of the lattices, and marking the edges of

every zone. These experimental pictures form by Bragg reflections from the regions near

the boundaries of the Brillouin zones of each lattice, resulting in dark stripes in the power

spectrum of the probe beam at the boundaries. We emphasize that the probe beam

intensity in these experiments is much lower than the intensity of the array-forming

waves (~1:10), and no external voltage is applied to the crystal while being probed.
Hence, Fig. 2 represents effects caused by linear propagation of the incoherent probe

beam in linear photonic lattices.

Next, we examine the spatial transmission spectrum of an incoherent probe beam

propagating in a hexagonal (trigonal) lattice with positive and negative defects. The

positive defect is created by launching an additional beam (several lattice periods wide),

which is co-propagating with the array-forming waves but is mutually incoherent with

them. Under a positive bias field, this additional beam induces a positive defect in the

lattice, i.e., it increases the index of refraction in the region it illuminates. Likewise, a

similar beam induces a negative defect when a negative bias field is applied. The upper

row in Fig. 3 (a-c) depicts the near-field photograph of the array-forming waves and the

defect-inducing beam as they exit the nonlinear crystal The lower row in Fig. 3 shows the

(far-field) power spectrum of the probe beam exiting the hexagonal lattice, without a

defect (d), with a positive defect (e), and with a negative defect (f) respectively. For the

positive defect, the guided modes (bound states) arise from total-internal-reflection of

states occupying the central region of the first Brillouin zone. Consequently, the far-field

of these guided modes is clearly apparent as a wide spot in the center of the first Brillouin

zone of Fig. 3(e). On the other hand, for a negative defect (for which the average

refractive index is lower in the guiding region), waveguiding arises solely from Bragg

reflections, with no contribution from total-internal-reflections. As a result, the central

region of the Brillouin-zone picture is empty [not populated; the central “hole” in Fig.

3(f)], and the guided modes are modes whose momentum arises from the vicinity of the

edge of the first BZ [bright concentric ring in Fig. 3(f))].

 Finally, when the photonic lattice is nonlinear, the underlying self-focusing (or self-

defocusing) interaction among the Bloch states results in energy exchange between these

lattice eigen-modes. This interaction facilitates a method for distinguishing between the

regions of normal and anomalous diffraction (dispersion) of the underlying lattice, and

mapping out the boundaries between dispersion of opposite signs. In these experiments,

the probe beam propagates nonlinearly in the photonic lattice, and it drives the nonlinear

interaction by inducing a broad defect in the lattice structure, which in turn causes energy

exchanges between Bloch states. To do this efficiently, the probe beam intensity must be

much higher than in the experiments described by Figs. 2 and 3. In this set of experiments

with nonlinear lattices, we use probe beam intensities which are roughly one-half of the

intensity of the lattice-forming array waves. Typical experimental results are shown in

Fig. 4. The excitation (far-field) power spectrum of the incoherent probe-beam is shown

in 4(a). When the crystal is biased with a positive field (self-focusing nonlinearity), the

wide probe beam induces a wide (and shallow) positive defect in the lattice (deeper

potential / increased refractive index). The presence of such a defect causes Bloch waves

from anomalous diffraction regions (negative curvature of the band structure) to transfer

power to Bloch waves arising from normal diffraction regions. Indeed, Fig. 4(b) clearly

reveals the regions of normal diffraction, which have a considerably higher intensity than

the low-intensity regions of anomalous diffraction. In a similar fashion, a negative bias

field results in a self-defocusing nonlinearity, through which the incoherent probe beam

induces a wide negative defect. This negative defect causes energy transfer from normal-

diffraction bloch waves to anomalous-diffraction states, highlighting the higher intensity

(anomalous diffraction) regions in Fig. 4c, adjacent to the lower intensity (normal

diffraction) regions in the same far-field power spectrum.

In conclusion, we have shown two different techniques for linear and nonlinear

Brillouin zone spectroscopy of photonic lattices, with or without defects. Both methods

rely on probing the lattice with an incoherent probe beam and visualizing the far-field

power spectrum of the light exiting the photonic structure. The linear method relies on

Bragg reflections, resulting in dark stripes along the edges of Brillouin zones, thus

marking the edges of every zone. The underlying linear mechanism arises from energy

transfer between waves propagating at the close vicinity of the Bragg angles (within the

Bragg mismatch), hence the preferential marking of the edges of the various Brillouin

zones. Our method of linear Brillouin zone spectroscopy is related to the Kikuchi

patterns (also known as Kossel lines) that have been observed in atomic structures

(crystals) through electron microscopy since the 1920's [17,18]. Similar features have

been observed in neutron scattering experiments, also in atomic lattices [19]. Finally,

recent experiments with stimulated Raman scattering of matter waves in Bose-Einstein

condensates (BEC) have also shown features of the extended Brillouin zone map of

atomic lattices [20]. However, in the BEC case, the experiments revealed the different

populations of the various bands (relying on different band energies), rather than marking

the borders between bands through Bragg diffractions (as in our experiments and in those

with the Kikuchi patterns). For this reason, the technique used in BEC cannot distinguish

between bands that are not separated by a complete gap [20]. Altogether, however, we
emphasize, that to the best of our knowledge no such methods for Brillouin zone

spectroscopy have ever been demonstrated in photonic structures.

In contrast to our linear Brillouin zone spectroscopy method (which has similarities to

in electron microscopy and neutron scattering), our nonlinear method adds new features

that have never been observed in any system. Namely, our nonlinear Brillouin zone

spectroscopy method marks the areas of normal and anomalous dispersion (diffraction;

effective mass, etc.) wherever they occur within the Brillouin zones of the lattice. This

method relies on the nonlinear propagation of the probe beam in the photonic lattice. The

probe induces (in real time) a broad defect in the lattice, which results in energy exchange

between Bloch waves. During a self-focusing interaction, Bloch waves from anomalous

diffraction regions transfer power to Bloch waves arising from normal diffraction

regions, and vice-versa for a self-defocusing interaction. The underlying physics behind

the nonlinear energy exchange among Bloch states has been intriguing researchers since

the famous Fermi-Pasta-Ulam era [11], and its fine details are still not fully understood.

Moreover, The ideas underlying our nonlinear Brillouin zone spectroscopy method most

probably cannot be introduced to electron microscopy or neutron scattering, simply

because the latter processes are fundamentally linear. In particular, it would be very

difficult to observe nonlinear interactions between electron (or neutron) waves in a

crystal and keep the crystalline structure intact. Nevertheless, our nonlinear technique

does have immediate relevance to other nonlinear periodic systems (e.g. dynamics in

BEC, plasma, sound, etc.), many of which are now being explored in the context of

“discrete” solitons [14].

 Finally, in addition to characterizing fully periodic photonic lattices, we also used

these techniques to characterize photonic lattices with positive and negative defects,

observing the far-field power spectrum of their guided and radiation modes. These

defect-bearing lattices are akin to photonic crystal fibers and related photonic structures.

We have demonstrated our techniques experimentally in optically-induced lattices;

nevertheless, the Brillouin zone spectroscopy method is general and can be used for any

periodic optical structure.

References

1. J. D. Joannopoulos, R. D. Meade, J. N. Winn, Photonic Crystals: Molding the

Flow of Light (Princeton Univ. Press, Princeton, 1995)
2. E. Yablonovitch, Phys. Rev. Lett. 58, 2059 (1987).
3. S. John, Phys. Rev. Lett., 58 2486 (1987)
4. J. C. Knight, J. Broeng, T. A. Birks, P. St. J. Russell, Science 282, 1476 (1998).
5. R. F. Cregan et al., Science 285, 1537 (1999).
6. B. Temelkuran et al., Nature 420, 650 (2002)
7. N.W. Ashcroft, N.D. Mermin, Solid State Physics (Saunders College,
Philadelphia, Pa.1976)
8. P. St. J. Russell, Appl. Phys. B 39, 231 (1986)
9. Y. Fink, et al., Science 282, 1679 (1998).
10. P. St. J. Russell, Science 299, 358 (2003).

11. O. Painter et al., Science 284, 1819 (1999).

12. H. S. Eisenberg, Y. Silberberg, R. Morandotti, A. R. Boyd, J. S. Aitchison, Phys.

Rev. Lett. 81, 3383 (1998).

13. J. W. Fleischer et al., Nature 422, 147 (2003).

14. D. N. Christodoulides, F. Lederer, Y. Silberberg, Nature 424, 817 ( 2003)

15. E. Fermi, J. Pasta, S. Ulam, Los Alamos Rpt. LA 1940 (1955).

16. J.W. Fleischer et al., Phys. Rev. Lett. 92, 123904 (2004)

17. S. Kikuchi, Jpn. J. Phys 5, 83 (1928).

18. W. Kossel et al., Z. Phys. 94, 139 (1935).

19. B. Sur et al., Phys. Rev. Lett. 88, 065505 (2002).

20. M. Greiner et al., Phys. Rev. Lett. 87, 160405 (2001); see also M. Greiner, PhD
thesis, Ludwig-Maximilians-Universität University, München, Germany, 2003.
Figures

Figure 1: (a) First and second Brillouin zones of a 2D square lattice with the high-

symmetry points (Γ, X, and M) marked with white dots. (b) First two bands of the

transmission spectrum of a 2D square lattice. (c) Dispersion curves between the

symmetry points of the first two bands. Negative curvature in these curves

corresponds to normal diffraction regions. (d) Diagram of our setup for Fourier-

space spectroscopy of photonic lattice.

Figure 2: Experimental linear mapping of the edges of Brillouin zones of square and

hexagonal lattices. (a) Interference pattern of the array waves forming the square

lattice at the crystal input face. (b) Fourier spectrum of the probe (broad circle) and

lattice forming beams (four sharp peaks) at the input face. (c) Calculated extended

BZ scheme of a square lattice. (d) True experimental picture depicting the Fourier

spectrum of the probe beam at the output face of the crystal with induced 2D

square lattice. (e)-(h) Same as a-d with a hexagonal lattice.

Figure 3: Defect modes in a hexagonal lattice. Interference pattern of the waves

forming the hexagonal lattice at the crystal input face (a) with no defect, (b) with

a positive defect, and (c) with a negative defect. Experimental Fourier spectrum of

the probe beam at the output face of the crystal with induced 2D hexagonal lattice.

(d) with no defect, (e) with a positive defect, and (f) with a negative defect.
Figure 4: Nonlinear dispersion mapping of a square lat1tice. (a) Fourier spectrum of

the probe beam (circle) and of the lattice forming beams (four dots) at the input.

Experimental Fourier spectrum of the probe beam at the output face of the crystal

with the induced 2D square lattice under (b) self-focusing nonlinearity and (c) self-

defocusing nonlinearity