1. Introduction

Organic electro-optic materials combine very fast nonlinearities with low dielectric constants which result in larger optical bandwidths of over 100 GHz compared to standard inorganic materials used in telecommunication[1]. The two organic material classes, electro-optic polymers and organic crystals have been intensively studied as integrated optics constituents during the last years. Polymers are potentially cheap and thin film processing is easy, nevertheless they often show thermal and photochemical instabilities and their nonlinearities are generally lower than in the crystalline counterparts. On the other hand, organic crystals have superior nonlin-earities, thermal and photochemical stability, but growth of high quality thin films is still a challenging step on the way to develop very large scale integrated (VLSI) photonic devices.

DAST (4-N, N-dimethylamino-4’-N’-methyl stilbazolium tosylate) is a widely investigated organic crystal with large electro-optic coefficients r ₁₁ = 77 ± 8pm/V at 800nm and r ₁₁ = 47 ± 8pm/V at 1535nm combined with a low dielectric constant (ε₁ = 5.2)[2, 3]. Therefore it is an interesting candidate for high speed applications in telecommunications provided that optical waveguiding is achievable.

Many techniques to produce waveguiding structures in DAST have already been proposed, including photobleaching[4, 5], photolithography[5], fs laser ablation[6], graphoepitaxial melt growth[7] and thin film solution growth[8, 9]. However, a fast standard technique allowing for a reliable thickness control of the guiding layer to a few tens of a micrometer, which is a prerequisite for the production of integrated optical structures, has not yet been established.

We decided to employ ion implantation since it is widely used for the production of waveguides in inorganic materials, for which other techniques can not be applied[10]. We present the fabrication of planar optical waveguides in an organic crystal by ion implantation. We show that the process of H⁺ and He⁺ ion implantation in organic crystals is essentially different from the phenomenon in inorganic crystals or polymers. We demonstrate that the thickness of the guiding layer can be controlled to a few tens of micrometers by H⁺ implantation. Furthermore, we present measurements of the refractive index profile and waveguide modes, and finally demonstrate planar waveguiding by end-fire coupling.

2. Ion implantation experiments

The crystals used in our experiments were grown in our laboratory from supersaturated methanol solution by the temperature lowering method[11]. DAST belongs to the monoclinic point group m[2]. The crystallographic b axis and the dielectric x ₂ axis are normal to the mirror plane, and the dielectric x ₁ axis makes an angle of 5.4° to the polar a axis in the symmetry plane[4]. The single crystals were cut approximately perpendicular to the dielectric axes. The typical sample size was around 7×4×4 mm³ with the longest dimension along x ₁. In order to use the largest electro-optic coefficient r ₁₁, the guided light should be polarized parallel to x ₁ and the refractive index n ₁ should be altered to ensure confinement. The x ₁ x ₂ surface polished to λ/4 surface quality was implanted at room temperature with He⁺ or H⁺ ions. The ion current was kept at about 2 nA and 40 nA for He⁺ and H⁺ ions, respectively, while the beam was scanned over an area of 9×9mm² to avoid thermal damage and to ensure homogenous implantation.

In order to get waveguides of small thickness we used the ions with the lowest energy achievable with the 6MeV EN Van de Graaff tandem accelerator, that is 720 keV He⁺ ions and 1MeV H⁺ ions. To avoid channeling along a crystallographic axis, the samples were tilted relative to normal incidence by 10° for He⁺ ion implantation. To additionally reduce the implantation depth of H+ ions the samples were tilted by 60°. This shortens the implantation depth by a factor of 2, having approximately the same effect as reducing the ion energy to 650 keV at normal incidence according to SRIM (The Stopping and Range of Ions in Matter) calculations (www.srim.org).

3. Refractive index profiles of He⁺ and H⁺ ion implantation

3.1. Model of ion implantation

The penetrating ion is loosing energy on its track in the target material. This energy deposition G(z) = dE(z)/dz, which can be calculated with SRIM, has been shown to be responsible for the refractive index change. G(z) is given by two contributions: electronic energy deposition G_el(z) = (dE(z)/dz)_el due to electronic excitations prevailing at high ion velocities, and nuclear energy deposition G_n(z) = (dE(z)/dz)_n due to collisions with nuclei that occur as the ion is slower at the end of the ion track. The index change induced by ion irradiation can be modeled assuming that the total refractive index change Δn(z) is a sum of the changes due to electronic and nuclear energy deposition[12]:

The phenomenological dependences are given by[12]

where Δn _{n, 0},Δn _{el, 0} are the saturation refractive index changes. Furthermore, G _n,0,G _{el, 0} represent the saturation energies and γ_n, γ_el the exponential factors for the nuclear and electronic energy deposition, respectively, while ϕ is the ion fluence. The parameter μ accounts for the limited accuracy in the prediction of the ion range obtained with SRIM.

3.2. Reflection scan measurement method

After implantation we investigated the optical properties of the samples by measuring the reflection of a wedge polished surface[13]. The refractive index n is related to the intensity reflection coefficient R = I_R/I ₀ by the Fresnel formula, where I ₀ is the incident light intensity and I_R the reflected one. For normal incidence and negligible absorption, n is given by

The sample was polished under an angle of 1° as seen in Fig. 1 to increase the spatial resolution of the measurement in z-direction. With a laser beam focused to less than 5μm the sample was then scanned in η direction in steps of 5μm yielding a depth resolution in z-direction of about 0.1μm. In order to probe the refractive index n ₁ the incident light was polarized parallel to the dielectric x ₁ axis that is normal to the ηz plane. The back reflected light was deflected by a beam splitter and then detected with a photodiode. Signal distortion due to polishing imperfections were reduced by averaging over 22 scans taken at different positions of the sample by translating it normal to the ηz-plane. In addition the back surface of the sample was polished under an angle of 8° to avoid collection of parasitic light (see Fig.1).

 

Fig. 1. Experimental setup for the determination of the refractive index profile by measuring the back reflected light from a wedged-polished sample, which was scanned in η direction in steps of 5 μm yielding a depth resolution of less than 0.1 μm in z direction.

Download Full Size | PDF

3.3. Refractive index profile of He+ ion implantation

First implantation experiments were done with 720 keV leading to an ion range of about 3.1μm in DAST as shown in Fig. 2. By comparing the magnitude of the electronic and the nuclear energy deposition and the area enclosed by the curves, we see that most of the ion energy is deposited by electronic excitations.

 

Fig. 2. Nuclear (blue) and electronic (red) energy deposition in dependence on the implantation depth for 720 keV He⁺ ions in DAST calculated with SRIM.

Download Full Size | PDF

 

Fig. 3. Refractive index change of n ₁ versus implantation depth z at the wavelength 633 nm for 720 keV He⁺ ion implantation in DAST with a fluence of ϕ = 0.8 × 10¹³ ions/cm² measured with the reflection scan method (Fig.1). The full line is a guide line for the eye.

Download Full Size | PDF

Implantation with a fluence of 1.5 × 10¹⁴ ions/cm² of 720 keV He⁺ resulted in a large reduction of the refractive index of around Δn ₁ ≈ -0.9 for λ = 633 nm at the surface. With a fluence reduced to 0.8 × 10¹³ ions/cm² a maximal refractive index change Δn ₁ of about -0.15 was measured at the surface as depicted in Fig. 3. In inorganic materials, such refractive index changes are only possible with fluences that are more than one order of magnitude higher. The refractive index profile in DAST after He⁺ implantation has a shape analogous to the electronic energy loss curve. Therefore, we can conclude that the refractive index is dominantly changed due to electronic energy deposition.

The effect of implantation is therefore essentially different compared to inorganic crystals, for which the nuclear displacements at the end of the ion track are the dominant mechanism of the refractive index change, leading to an optical barrier with reduced refractive index[10, 14]. The effect of implantation is also not the same as in the polymer PMMA[14, 15], where an increase of the refractive index was detected and attributed to an increased absorption. In He⁺ implanted DAST the refractive index was decreased. We attribute the lowered refractive index to a decrease of the polarizability of some molecules due to broken bonds.

We can conclude that He⁺ ion implantation is not suitable for the production of planar waveguide structures in organic crystals, because of the diffusion-like shape of the refractive index profile and the fact that the refractive index is reduced, resulting in an implanted layer with a lower index than the bulk substrate.

3.4. Refractive index profile of H⁺ ion implantation

Since the He⁺ experiments with ions showed that a decrease of the refractive index induced by electronic energy deposition in DAST is present, we selected an ion with appropriate electronic energy loss profile to produce a refractive index barrier. The electronic energy deposition curves for He+ and H⁺ ions differ considerably in DAST, as it is evident by comparing Fig. 2 and Fig. 4, respectively. Whereas for He⁺ ions it is similar to a diffusion-like profile with a maximum at the sample surface, the electronic deposition curve for H⁺ ions has a similar shape as the nuclear energy deposition curve showing a peak at the end of the ion track, which would result in a refractive index barrier.

 

Fig. 4. Nuclear (blue) and electronic (red) energy deposition curve of 450 keV (dashed line) and 1 MeV H⁺ ions (full line) in DAST calculated with SRIM. For reason of comparison, ordinate axes with the same order of magnitude as in Fig. 2 have been chosen.

Download Full Size | PDF

We employed a fluence of 1.25 × 10¹⁴ ions/cm² to produce a refractive index barrier. The measured profile of the refractive index change Δn ₁ is shown in Fig. 5 for wavelengths of 633 nm and 810 nm. Peak refractive index changes of around -0.2 at 633nm (n ₁ = 2.69) and -0.1 at 810nm (n ₁ = 2.36) were measured. The difference of these values in this small wavelength range can be attributed to the large dispersion of DAST since the wavelength of 633 nm is close to the absorption region. The solid curves in Fig. 5a) and b) are theoretical curves based on the model described in section 3.1 with the following remarks. According to the results of He+ implantation in DAST we can conclude that Δn_el » Δn_n and therefore the nuclear contribution to the total index change can be neglected. Since for low fluences far below the saturation level the exponential term of Δn_el(z) can be linearized, the parameters Δn _{el, 0} and G _{el, 0} cannot be independently modeled. Analysis showed that we are in the low fluence regime and therefore G _{el 0} was fixed to 50 × 10²² eV/cm³. With this assumption we determined the parameters of Eq.(3) which correspond best to our experimental data. The results are listed in Table 1. In the low fluence regime the exponential factor describes the nonlinear relation between the refractive index change and deposited energy. Therefore, values of γ_el above 1 are desirable for the production of optical waveguides in order to have a clear distinction between the barrier region and the guiding layer combined with small refractive index changes in the waveguide core.

 

Fig. 5. Refractive index change of n ₁ versus implantation depth z at the wavelength 633 nm (a) and 810 nm (b) for 1MeV H⁺ ion implantation in DAST with a fluence of ϕ=1.25×10¹⁴ ions/cm² measured with the reflection scan measurement method. The full lines are theoretical curves based on the theoretical model. The corresponding parameters are summarized in Table1.

Download Full Size | PDF

Table 1. Model parameters of Eq.(3), which correspond best to the experimental data shown in Fig. 5 obtained by a least square theoretical analysis.

View Table

4. Mode analysis

4.1. Barrier coupling method

In order to study the impact of the fluence on the refractive index profile, which is of great importance for a suitable waveguide tailoring, we employed implantations with three different fluences: 1.25×10¹⁴ ions/cm², 1.0×10¹⁴ ions/cm² and 0.5×10¹⁴ ions/cm². The mode profiles were measured with the recently developed barrier coupling method[16], which is more suitable for our organic crystal compared to prism coupling since no prism is required and the modes can be measured in a non contact configuration without applying pressure to the sample. A schematic illustration of the method is shown in Fig.6a). The method is based on frustrated total internal reflection on the waveguide barrier. The light beam with a vacuum wavenumber k ₀ is coupled from the substrate with refractive index n_s to the waveguide core through the optical barrier. If the component of the wave-vector along the waveguide propagation direction x ₂ of the beam β = k ₀√n_s ² -sin ²(θ) matches one of the waveguide mode propagation constants β = k ₀ ∙ N _eff, then the evanescent field of the reflected laser light is coupled to a waveguide mode. Coupling results in a dip in the reflection intensity. A measurement of the reflected light as a function of the external incident angle θ for a DAST sample implanted with 1 MeV H ⁺ ions with a fluence of 1.0 × 10¹⁴ ions/cm² is shown in Fig. 6b). Eight TE modes for laser light polarized along the dielectric x ₁ axis were detected. This measurement confirmed the creation of a waveguide by H₊ ion implantation.

 

Fig. 6. a) Experimental setup for the determination of the effective mode indices with the barrier coupling method. b) Detected reflectivity as a function of the external angle θat λ= 810 nm for light polarized along the dielectric x ₁ axis. The DAST sample was implanted with 1MeV H⁺ ions with a fluence of 1.0×10¹⁴ ions/cm².

Download Full Size | PDF

4.2. Results

We compared the effective mode indices measured by the barrier coupling method with those stemming from the refractive index profiles obtained in section 3.4 by the reflection scan method. These were calculated using a 2 × 2 matrix formalism[10]. The refractive index profiles for the three different fluences 1.25 × 10¹⁴ ions/cm², 1.0 × 10¹⁴ ions/cm² and 0.5 × 10¹⁴ ions/cm² obtained with the parameter summarized in Table 1 are shown in Fig. 7a) and the corresponding effective mode indices in Fig.7b). The measured modes (∘) and the calculated modes (full line) are in good agreement within the experimental error. The barrier of the produced waveguides is thick, which is good to keep the losses low. However, the barrier coupling method cannot be used for modes with low tunneling losses and therefore the first two modes were not detectable as is visible by comparing the measured and calculated modes. Beacuse of limited accuracy in the ion range predicted by SRIM, we varied the barrier position obtained by the reflection scan method (μ = 1.07± 0.05) within the experimental error. The dashed lines shown in Fig. 7a) correspond the best to our experimental data and were obtained by assuming a reduced implantation depth of about 10% (μ= 1.11) compared to SRIM calculations.

 

Fig. 7. a) Refractive index profiles atλ= 850 nm for fluences of 1.25×10¹⁴ ions/cm² (red), 1.0×10¹⁴ ions/cm² (blue) and 0.5×10¹⁴ ions/cm² (green) calculated with the parameters obtained with the reflection scan measurement method. b) Corresponding calculated (full line) and measured (∘) effective mode indices N _eff; the calculated modes indices are connected for clarity reasons. The dashed lines for the effective indices correspond best to the measured modes obtained by a reduction of the barrier position of about 5% compared to the depicted refractive index profiles in a).

Download Full Size | PDF

5. Waveguiding experiments

After implantation the x ₁ x ₃ end-faces were polished and waveguiding experiments performed by standard end-fire coupling. The light was propagating along the x ₂-direction and was polarized parallel to the dielectric x ₁ axis, which is the most interesting configuration for optical modulation in DAST in order to use the largest electro-optic coefficient r ₁₁. The experimental setup is shown in Fig.8a). Waveguiding was clearly observed at wavelengths of 840 nm,1315 nm and 1570 nm. Figure 8b) shows a photograph of the end-face as taken with the infrared camera of the setup with light coupled at 1570 nm.

 

Fig. 8. a) Experimental setup for the determination of the waveguiding characteristics and loss determination: L: lens, CCD: infrared camera, PD: photodiode. b) Photograph of the end-face of the crystal taken with the CCD camera. The cross section of the planar guided light at 1570 nm is clearly visible.

Download Full Size | PDF

The losses of the planar waveguide structures were determined by measuring the power of the guided light transmitted through the waveguide with the setup shown in Fig.8 at λ = 1.57μm. The guided light was collected with two lenses and detected with a photodiode. Special care was taken to only detect the guided light. Therefore, the polished end face of the crystal sample was imaged onto a horizontal slit that was aligned so that only the guided light could be transmitted. For an assumed coupling efficiency of 80%, the losses were 7±2dB/cm for the waveguide produced by the ion fluence of 1.0 × 10¹⁴ ions/cm² and about 10 dB/cm for two other waveguides. The contribution of bulk material absorption is about 2.5 dB/cm at 1.57μm. At the present state it is not clear if the slightly increased losses for lower 0.5 ×10¹⁴ ions/cm² and higher 1.25 × 10¹⁴ ions/cm² fluences are related to higher tunneling losses or defect concentration or to a reduced edge quality and therefore lower coupling efficiency.

6. Conclusions

In conclusion, we have realized optical waveguides in the organic crystal DAST by H⁺ ion implantation. Waveguiding with losses of 7 dB/cm at 1.57μm has been achieved in planar waveguides. We have shown that the main refractive index changes in the organic crystal DAST are due to changes in the molecular structure and are caused by the electronic excitations. This result is a new and unexpected finding and completely different from inorganic materials, for which the origins of the refractive index change are ion induced nuclear displacements. Therefore the work gives new perspectives for the use of ion implantation in integrated optics, being for the first time successfully applied to organic crystals. Refractive index changes in the barrier region of up to -0.1 at 810nm have been observed at relatively low fluences of the order of 10¹⁴ ions/cm². We have determined the refractive index profile as a function of the H⁺ ion fluence by measuring the reflectivity of a wedged-polished sample and confirmed by measuring the effective mode indices by the barrier coupling method. Using these data the effect of ion implantation could be accurately described with a model. Since the waveguide thickness can be tailored by the implantation process within a few tens of micrometers, an important first step towards integrated optics in DAST has been made.