High Optical Quality Organic Films for Third Order Nonlinear Optics

The main challenge to using organic materials for nonlinear optics is not just the creation of molecules that have a large third-order polarizability, but it is also to assemble these molecules into a solid state material that reflects the nonlinearity of the molecules and at the same time has a high optical quality. To achieve a high third-order susceptibility, one needs to pack a large number of molecules per unit volume, and one needs to have a molecular third-order polarizability that is large compared to the size of the molecule. To achieve a high optical quality one needs to create a homogenous material without any microcrystals or molecular clusters that can lead to light scattering. Our recent work on both topics has allowed us to develop a supramolecular material that is an amorphous dense assembly of small molecules with a high third-order nonlinearity and that can be efficiently combined with nanostructured silicon-on-insulator waveguides.

Bweh Esembeson with his high-vacuum machine.

As a continuation and spin-off from our research on organic molecules for nonlinear optics, we have investigated the use of high-vacuum molecular beam deposition of optimized small molecules to create organic thin films for integrated nonlinear optics. After looking at many molecules and studying the correlation between molecular design and thin film morphology we finally identified a good compromise between molecular structure and molecular nonlinearity that leads to films that have both a very high optical quality and a high bulk third-order susceptibility.

The figure below shows the DDMEBT molecule. This molecule is part of a family of molecules that we recently developed with the aim of optimizing the third-order optical nonlinearity of a molecular assembly (see here). It was synthesized in the Group of Prof. Diederich at ETH Zurich, Switzerland. We explicitly focused on molecules that are small and robust, and many of them can be sublimated without decomposition. This enabled the use of molecular beam deposition in high vacuum to create thin films with a large third-order nonlinearity. In the DDMEBT molecule, we merged the important molecular elements responsible for the high third-order nonlinearity with a molecular structure that, by decreasing intermolecular interactions, lead to a high quality supramolecular assembly. It is the non-planar geometry of DDMEBT, visible in the figure below, that promotes the formation -- upon vapor deposition -- of an amorphous supramolecular assembly without any grains or inhomogeneities.

Two related molecules with large intrinsic third-order susceptibilities. The TDMEE molecule has a planar structure. The DDMEBT molecule contains almost the whole structure of TDMEE as a subunit, but has an additional component that gives it a nonplanar structure. A three-dimensional view of DDMEBT is shown on the right.

DDMEBT films have a wide transparency range starting at wavelengths above 700 nm, with a modulation in the transmission at longer wavelengths that is a result of multiple reflections inside the film and that perfectly matches the predicted modulation for a featureless, flat, transparent film. The existence of such a clear transmission modulation with wavelength already means that any thickness variations within an area of several millimeters squared must be much smaller than the wavelength of the light inside the material, i.e. smaller than a few tens of nanometers. The homogeneity of the DDMEBT films is confirmed by surface reliefs obtained by Atomic Force Microscopy on a 950 nm thick film. Height fluctuations were less less than 5 nm over distances of 0.5 micrometers.

The third-order nonlinear optical susceptibility of the DDMEBT films is off-resonant at telecommunication wavelengths. Degenerate four-wave mixing (DFWM) experiments at 1.5 micrometers found it to be 1000 times larger than that of fused silica.

Transmission spectra of three DDMEBT films with thicknesses (beginning from top) of 380 nm, 450 nm and 950 nm, respectively. The data are represented by white circles and are shifted vertically for clarity. The thick solid curves are the results of a simultaneous least squares fit with the refractive index dispersion shown in the inset and with no other free parameters. The refractive index at 1.5 micrometers is 1.8

We can compare the third order nonlinearity of the DDMEBT supramolecular assembly with what is probably the organic material with the highest off-resonant third-order susceptibility demonstrated to date, the PTS single crystal. PTS has a nonlinear refractive index 10 times larger than that of the DDMEBT material, but this high value can be observed only for one particular polarization direction of the light corresponding to the alignment of the chromophores in the crystal. The same density of the same chromophores, but randomly oriented like the DDMEBT molecules in our film, would produce a 5 times smaller third-order susceptibility. Another important point is that fabrication of extended single-crystal thin films of PTS is quite difficult, and the inherent anisotropy of the material is an important handicap when it comes to integration with existing photonic devices. In contrast, by vapor deposition of DDMEBT we can easily cover a wide area (many square centimeters) of any substrate with a high quality nonlinear optical film that has a large isotropic third-order susceptibility.

It is important to note that the combination of such a large off-resonant third-order susceptibility (three orders of magnitude larger than fused silica) with a small refractive index (n ~ 1.8) is a unique property of an organic material that enables its use as the active cladding in specially designed silicon waveguides. As an example, third-order susceptibilities of similar magnitude may be found in some chalcogenide glasses, but they are accompanied by a large refractive index that does not support the optical field enhancement characteristic of SOH structures.

As a conclusion, we have developed a new molecular material that effectively combines a high optical quality with a wide deposition area and a large, isotropic, off-resonant third-order susceptibility that is three orders of magnitude larger than that of fused silica at telecommunication wavelengths. In particular, we have shown that vapor deposition of this organic material in high vacuum makes it possible to perfectly and homogeneously fill the less than 200 nm wide slot between two silicon ridges where the optical mode is concentrated in a silicon-organic-hybrid waveguide.

The DDMEBT supramolecular assembly represents an important breakthrough in organic nonlinear optics, with numerous opportunities for realizing all-optical switching devices in integrated optics.

Left: AFM scan of the edge of a DDMEBT film, showing the transition from a part of the substrate that was protected by a shutter during deposition, to the exposed part where the film was deposited. Middle: SEM picture of the film, showing the absence of any grains down to the nanoscale. Right: Spectrum of the molar extinction coefficient of DDMEBT in a dichloromethane solution and of the transmission of a 950 nm thick DDMEBT thin film.

Atomic force microscope scans showing surface-relief in a 950 nm DDMEBT film over an area of 500 nanometers squared. The figure on the left shows the full surface-relief as a color image. The figure on the right is a plot of 5 line scans taken every 100 nm, with the data displaced vertically by 5 nm for clarity