Journal of Sol-Gel Science and Technology

, 48:80

Waveguides and gratings fabrication in zirconium-based organic/inorganic hybrids


  • Carlos M. S. Vicente
    • Department of Physics and Instituto de TelecomunicaçõesUniversity of Aveiro
  • Edison Pecoraro
    • Department of Physics and Instituto de TelecomunicaçõesUniversity of Aveiro
  • Rute A. S. Ferreira
    • Department of PhysicsCICECO, University of Aveiro
  • Paulo S. André
    • Department of Physics and Instituto de TelecomunicaçõesUniversity of Aveiro
  • Rogério Nogueira
    • Department of Physics and Instituto de TelecomunicaçõesUniversity of Aveiro
  • Younes Messaddeq
    • Institute of ChemistrySão Paulo State University, UNESP
    • Institute of ChemistrySão Paulo State University, UNESP
    • Department of PhysicsCICECO, University of Aveiro
Original Paper

DOI: 10.1007/s10971-008-1775-3

Cite this article as:
Vicente, C.M.S., Pecoraro, E., Ferreira, R.A.S. et al. J Sol-Gel Sci Technol (2008) 48: 80. doi:10.1007/s10971-008-1775-3


Sol–gel derived poly(oxyethylene)/siloxane organic/inorganic di-ureasil hybrids containing different amounts of methacrylic acid (McOH, CH2=C(CH3)COOH)) modified zirconium oxo-clusters (Zr-OMc) were processed as thin films deposited in glassy substrates via spin coating and as transparent and shape controlled monoliths. Channel monomode waveguides and diffraction gratings were UV patterned using the Talbot interferometer and the Lloyd mirror interferometer experimental setups. The time dependence of the diffraction gratings efficiency was studied for hybrids containing different amounts of Zr-OMc. Finally, the number of propagating modes and the refractive index gradient within the waveguide region, determined as a Gaussian section located below the patterned channel, was evaluated and modeled, a maximum index contrast of 2.43 × 10−5 being estimated.


Organic/inorganic hybridsIntegrated opticsSol–gelWaveguidesZirconium oxo-clusters

1 Introduction

During the last years, the fabrication of integrated optics devices using hybrid materials has received an increasing amount of attention [15]. The advantages of the sol–gel process, such as the low-temperature processing and shaping, high sample homogeneity and purity, the possibility of mixing the inorganic and organic precursor components at the nanometer scale, and the availability and low cost of numerous metallo-organic precursors makes this method very suitable for the development of organic/inorganic hybrid materials for the production of functional integrated optics devices [6, 7]. Among the various organic/inorganic hosts that have been developed in the last years, those containing amine functionalities, namely urea cross-linked hybrids classed as di-ureasils [811], present acceptable transparency, mechanical flexibility and thermal stability to be processed as thin films [1214]. The combination of di-ureasils with metal oxide precursors, specifically methacrylic acid (McOH, CH2=C(CH3)COOH)) modified zirconium tetrapropoxide, Zr(OPrn)4, is recursively used for the fabrication of integrated optics devices. Recently, the local structure of these di-ureasil organic/inorganic hybrids was discussed in terms of the Zr-OMc content [15]. For low Zr-OMc content the Si- and Zr-based networks are inter-constrained, since the Zr-based clusters are embedded in the polymeric phase between the siliceous domains, whereas segregation of the individual components at the 0.1 μm scale occurs for high contents [15]. The presence of increasing size Zr-rich aggregates (∼200–450 nm) as the Zr-OMc relative content in the hybrids increases was detected [15]. It was also demonstrated that the chemical process that takes place under UV illumination is the polymerization of the methacrylate groups of the Zr-OMc aggregates [15].

Di-ureasil hybrids have already been used as integrated optics substrates, namely in the production of patternable gratings, channels and monomode planar waveguides with low propagation losses (<0.3 dB/cm), presenting good thermal stability [12, 13, 15]. Distributed feedback lasers (DFB) by using dynamic gratings have been demonstrated for di-ureasil thin films incorporating rodhamine 6G [16]. Furthermore, Fabry-Perot cavities were also obtained by UV writing without the need of photoinitiators revealing a reflection coeficient of 0.042 with a free spectral range (FSR) value of 35.6 GHz which makes it adequate to be used in optical clock recovery in optical signals with bit rate of 35.6 Gbit/s [15].

In the context of the development of innovative integrated optics devices, such as low-cost optical power splitters for the general use spreading of all optical access networks a window of opportunities is open for sol–gel derived organic/inorganic hybrids. Examples of applications encompass narrow band optical filters (used as demultiplexers to access the desirable wavelengths in a multi-wavelength system), low losses optical power splitters, and optical cavities (for the optical clock extraction function).

In this work, we use di-ureasil-zirconium oxo-clusters organic/inorganic hybrids processed as thin films and transparent monoliths as integrated optics substrates. UV patterned channel monomode waveguides and diffraction gratings are obtained without the need of photoinitiators. In particular, diffraction gratings have been written inside and outside the channel waveguides and the refractive index contrast will be compared. The diffraction gratings efficiency was studied as function of time and Zr-OMc concentration. The mode field distribution within the waveguide region, previously set as a Gaussian section located below the patterned channel [15], is addressed in greater detail and the refractive index contrast modeled.

2 Experimental

2.1 Materials synthesis and processing

2.1.1 Di-ureasil precursor (d-UPTES(600))

The organic–inorganic hybrid framework, designated as d-U(600), is composed of polyether-based chains grafted at both ends to a siloxane backbone through urea functionalities [810, 12, 15], whose molecular structure is depicted in Scheme 1.
Scheme 1

Molecular structure of the non-hydrolyzed di-ureasil d-U(600) precursor (a + c = 4.5, and n = 8.5)

2.1.2 McOH-modified Zr(OPrn)4 precursor

The precursor used in the preparation of the hybrid materials, named hereafter as Zr-OMc, together with d-UPTES(600), was obtained by mixing Zr(OPrn)4 (Fluka) and McOH at room temperature with a molar ratio Zr(OPrn)4:McOH=1:1 in butanol (CH3(CH2)3OH, BuOH) and stirred for 3 h [15].

2.1.3 Di-ureasil-Zr-OMc hybrids

The Zr-OMc and d-UPTES(600) precursors were pre-hydrolyzed with a hydrochloric acid solution (HCl 0.01 mol L−1), with a water-to-metal molar ratio of 0.5:4. The pre-hydrolyzed solutions were mixed with different Zr relative molar percentage ranging from 5% to 75%. The synthesis details were given elsewhere [12, 15].

The di-ureasil-Zr-OMc hybrids is henceforth identified as d-UZ(X mol% Zr) and were processed both as transparent and with shape control monoliths and as thin films on glass substrates using the spin-coating technique.

2.1.4 Fabrication of monomode waveguides and gratings

The channel waveguide writing was performed in the monoliths through the exposure to a c.w. Ar-ion laser, frequency doubled (244 nm) by a BBO crystal, operating at 40 mW. The laser beam was shaped by an iris in order to remove spatial noise and by a plano-cylindrical lens to create a narrow laser line with 2 mm length. Then, the laser line was translated by a motorized positioning system in order to create the channel waveguide. The speed of the translation and the exposure time were varied between 0.02–0.08 mm/s and 2–18 min, respectively.

Diffraction gratings were recorded under different methods and laser radiations. One method is based on a modified Talbot interferometer with a phase mask acting as beam splitter, using the laser source described above for the waveguide patterning [15]. This method was used for the d-UZ(23) monolithic hybrids.

The main advantage of this method resides in the easier alteration of the focal point and therefore of the interference conditions that result in a change of the interference period, giving extra tunability. In the second method, applied for d-UZ(X) thin films (X = 15, 50 and 85), a Lloyd mirror interferometer set-up was used. The UV line of a Kripton laser 337.5 nm and 400.0 mW was focused on the sample with a cylindrical lens to a 20 × 2,000 μm stripe. One part of the incident beam reaches directly the film while the other part overlaps the first one after reflection on the mirror. An interference pattern is created on the sample. The grating period, Λ, corresponding to the space between fringes in the interference pattern is given by:
$$ \Uplambda = \frac{{\lambda _{1} }} {{2\sin \alpha }} $$
where λ1 is the laser wavelength and α is the angle between the incident beam axis and the mirror plan. α was varied in order to achieve different gratting periods, enabling the tuning of the difraction spectral band. A He-Ne laser was used to measure the diffraction efficiency in real time. Diffraction efficiency was defined as the ratio between the intensities of the 1st and zero order diffractions.

2.2 Characterization

2.2.1 Atomic force microscopy (AFM)

The images were obtained using a AFM Nanoscope Instruments equipment, in tapping mode, with a super sharp silicon probe having a radius of 10 nm, resonance frequency 330 kHz and spring constant 42 N/m. The images were deconvoluted considering the probe’s shape using the software WSXM® [17]. In order to improve the images quality, flattening and elimination of line noise tools and a Gaussian filter were used. The same tip was employed in all the images recording to avoid the influence of the tip radius variations in the square roughness values.

2.2.2 Mode field distribution

The modal field distribution was acquired using a positioning system and a Laser Beam Profiler from Newport LPB-1 with microscope objective. The measurements were made with two lasers operating at 632 nm and 980 nm. The light was coupled into the waveguide region previously patterned, using an optical fiber and a xyz positioning system (Thorlabs Namomax) in order to excite a propagation mode. The propagating region limits were set when the output light tends to zero. The output mode profile was measured with a beam profile analyzer using an integration time of 10 s and the xy position experimental error is 0.15 μm.

2.2.3 Refractive spectrum

The refractive grating spectrum of the monolithic sample was obtained with an Amonics ALS-CL-17optical broadband source covering the spectral region between 1,530 nm and 1,620 nm. The signal is coupled to the waveguide using a standard monomode optical fiber. In order to minimize de optical losses in the coupling, an index matching gel (Thorlabs) was used. An optical circulator monitored the reflected spectrum, measured with an optical spectrum analyzer (Advantest Q8384).

3 Results and discussion

Monomode waveguides were written, without the need of photoinitiators, through the exposure of the hybrid monoliths to UV light under different laser line speed and time interval. Figure 1a shows the UV exposed surface of the d-UZ(23) hybrid that presents acceptable homogeneity [15]. The UV patterned hybrids’ surface displays a material contraction with a Gaussian shape, arising from the laser beam spectral distribution, as previously reported [15]. The cross section profile is plotted in Fig. 1b with dimensions (wide × deep) of 50 μm × 110 nm. The channel dimension depends on the laser characteristics. Indeed, the channel presented is shallower than those previously reported using the same laser rate translation and exposure time and higher laser power (40 mW). The contraction of the surface, shown in Fig. 1, is interpreted as resulting from the local UV-induced polymerization, allowing a densification of the hybrid underneath the exposed area, inducing an increase in the refractive index, relatively to that of the unexposed area [15].
Fig. 1

(a) AFM images and (b) cross section profile of the channel waveguide written in the d-UZ(20) hybrid (laser speed and time exposure of 0.05 mm/s and 7 min, respectively)

Two diffraction gratings were superimposed in this channel, Fig. 2a, b. The grating pitch values are estimated by fitting the experimental AFM data (Fig. 2a, b) to a sinusoidal function (not shown), revealing periodic refractive index perturbations of Λ = 470 and Λ = 530 nm, respectively, with a grating relief depth around 3 nm. The different grating pitch values were achieved by changing the experimental interferometer conditions, namely the angle between the incident beam axis and the mirror plan (Eq. 1). The AFM images either of the channel waveguide and of the diffraction gratings evidence low root mean square roughness values (<1 nm).
Fig. 2

AFM images of two diffraction gratings written in the d-UZ(23) hybrid superimposed on the channel waveguide previously patterned (laser speed and time exposure of 0.05 mm/s and 7 min, respectively)

The diffraction efficiency of the gratings was evaluated in thin films as function of the time exposure to the beam in the Lloyd interferometer for different Zr-OMc concentrations. As shown in Fig. 3, with the increasing of the Zr-OMc relative content faster grating formation was observed, the UV-induced refractive index change is proportional to the amount of Zr-OMc concentration. The maximum efficiency converges to ∼1.5%, independently of the Zr-OMc amount.
Fig. 3

Exposition time dependence of the diffraction efficiency in the hybrids (a) d-UZ(85), (b) d-UZ(50) and (c) d-UZ(15) written using the Lloyd mirror interferometer set-up

The light propagation was quantified through the estimation of the effective propagating region and refractive index, using the experimental procedure previously adopted [15]. The guidance properties of a UV written channel in the d-UZ(23) monolith were further studied acquiring the modal field distribution for propagation in the visible and infrared spectral regions. The position of the injected signal was varied along the width and height of the propagating region, exciting other transversal modes, whose frontiers were set when the output light tend to zero. From this experiment the guidance section was determined as a Gaussian region placed below the patterned channel, with a typical width of 280 μm (Fig. 4a). The mode field distribution at different locations within the propagating region is shown in Fig. 4b–f. Loyd interference pattern is observed in the mode field distribution acquired closely to the surface. The channel centre displays monomode type propagation, whereas moving away towards the limits of the propagating region two propagation modes were discerned, indicating a gradient in the refractive index contrast.
Fig. 4

(a) Scheme of the waveguide region in the d-UZ(33) monolith. Mode field of the waveguides excited at 632.8 nm at different locations within the guidance region: (b) channel centre; (c), (d) along the y-axis, and (e), (f) along the x-axis. The (x, y) coordinates within the guidance region are set bellow each figure in μm

With the goal of quantifying this gradient, the mode field distribution was modeled through a near-field method to enable the reconstruction of the index profile. This method is based in the relation between the refractive index profile, n(x,y), and the electrical field, E(x,y), distribution. For a guided mode the following relation must be satisfied [18]:
$$ \nabla ^{2} E{\left( {x,y} \right)} + {\left[ {k^{2} n^{2} {\left( {x,y} \right)} - \beta ^{2} } \right]}E{\left( {x,y} \right)} = 0 $$
where k is the free space wavenumber and β is the mode propagation constant. For a small refractive index contrast, Δn(x,y),
$$ n^{2} {\left( {x,y} \right)} = n^{2}_{s} + 2n_{s} \Updelta n{\left( {x,y} \right)} $$
being ns the substrate refractive index. The index profile is represented as a function of the measured near field intensity, I(x,y) = (E(x,y))2, allowing the determination of the k and β values:
$$ \Updelta n{\left( {x,y} \right)} = {\left[ {\frac{{\nabla ^{2} {\sqrt {I{\left( {x,y} \right)}} }}} {{2kn_{s} {\sqrt {I{\left( {x,y} \right)}} }}} - \frac{{\beta ^{2} }} {{2kn_{s} }}} \right]} - \frac{{n_{s} }} {2}. $$
The electrical field spatial distribution was obtained from the square root of the intensity of the mode field distribution (Fig. 4). The elements of the intensity matrix were smoothed by convoluting the signal with a Gaussian filter, with a bandwidth of 40 pixels, resulting in the field distribution displayed in Fig. 5. At this stage the refractive index contrast can be obtained from the direct application of Eq. 4 to the electric field matrix values. The sensitivity of this method is limited to the mathematical operations involved requiring the application of a Gaussian filter to smooth the calculated index step. Figure 6 shows the index contrast in the waveguiding region, evidencing a gradient of the refractive index values in that region, being higher in the center of the channel.
Fig. 5

Electrical field distribution using the intensity field matrix smoothed by convolving the signal with a Gaussian filter
Fig. 6

Refractive index contrast obtained from Eq. 4

The region where the refractive index value undergoes a modification can be estimated by the limits where the contrast tends to zero, yielding to dimensions of 300 μm × 200 μm at 1/e4, which is comparable with the previously estimated area (Fig. 4a). A maximum refractive index contrast of 2.43 × 10−5 was also estimated, enabling the guidance in the center of the UV exposed region. Such a low refractive index contrast is in good agreement with experimental results. The refractive index measured after UV exposure within the guidance region has been determined as 1.5162 [15], which is very close to the values reported for unexposed organic/inorganic hybrids with approximately the same Zr-OMc content (1.5133), [12]: pointing out a small refractive index contrast induced by the UV exposure as predicted here. To accurately quantify the central waveguiding region where the maximum value for the contrast was determined and monomode propagation occurs, the variation of the refractive index values along the vertical and horizontal direction in Fig. 6 were analyzed. Figure 7a, b shows the vertical and horizontal refractive index profiles, respectively. The vertical profile is characterized by an intense peak near the surface plane, corresponding to the focal point of the UV laser, followed by a broad peak corresponding to deeper regions, resulting from a secondary focus point due to aberrations on the UV writing system optics. The horizontal profile shows a central intense peak followed by two lateral peaks also due to aberrations and ghost images on the writing system optics. The fit of the main peaks to Gaussian functions allows us to estimate the 1/e width for the monomode waveguide region, yielding to 38.8 μm and 63.8 μm for the vertical and horizontal directions, respectively.
Fig. 7

Refractive index contrast profiles along the (a) vertical and (b) horizontal directions

4 Conclusion

Channel monomode waveguides were UV written in homogeneous transparent poly(oxyethylene)/siloxane di-ureasils containing McOH modified Zr(OMc) processed as monoliths and thin films. The Talbot interferometer and the Lloyd mirror interferometer experimental setups were successfully used to write difraction gratings, either in monolithic samples surfaces or thin films deposited on glass substrates. The maximum diffraction efficiency was observed to be independently of the Zr-OMc relative content. On the other hand, the higher is the Zr-OMc content the faster is the patterning. The patterned samples were also characterized by AFM. Very low roughness values below 1 nm were evaluated. From mode field distribution experiments the guidance section was determined as a Gaussian region placed below the patterned channel, with a typical width of 280 μm and a maximum refractive index contrast of 2.43 × 10−5 in close agreement with previously reported experimental data. The dimensions of the central waveguide region with higher refractive index contrast and monomode propagation were estimated as 38.8 μm × 63.8 μm, through the analysis of the variation of the refractive index values along the vertical and horizontal directions.


The support of NoE “Functionalised Advanced Materials Engineering of Hybrids and Ceramics” (FAME) is gratefully acknowledged. This work was also supported by Fundação para a Ciência e Tecnologia (Portuguese agency), FEDER and POCI programs under contract POCI/CTM/59075/2004, PTDC/CTM/72093/2006, FAPESP and CNPq (Brazilian agencies) and CAPES-GRICES Brazil–Portugal cooperation program, contract BEX 2866/05-6. The authors thank the help of Daniela C. Oliveira (LNLS), Máximo S. Li (USP) and Andreia G. Macedo (University of Aveiro) for some synthesis, diffraction efficiency measurements and AFM analysis, respectively.

Copyright information

© Springer Science+Business Media, LLC 2008