Novel thin films target device applications

Finally, Sir Isaac Newton of the UK defined a new optical constant, the refractive index, that allowed him to convert the proportionalities found by Kepler and Snell into an easy-to-use equation. In addition to refraction, the refractive index is a critical quantity for other optical phenomena, such as Fresnel reflection, wave propagation and wave guiding.

Figure 1

Despite the importance of the refractive index, no conventional thin-film materials exist with a refractive index between 1.0 and 1.4 (figure 1a). Yet the availability of materials within this range would be extremely desirable, since the performance of many photonic components could be strongly enhanced if such low refractive-index materials were available.

Refractive index at the extreme
To address this issue, we have developed a thin-film deposition technique, known as oblique angle evaporation, that yields optical thin films with a refractive index as low as 1.05. In this technique, the sample surface is positioned at an angle to the incoming physical vapour flux. The result is a self-organizing nanostructure that consists of an array of inclined nanorods (figure 1b).

Statistical fluctuations in the height of the film surface produce shadow regions that the incoming vapour flux cannot reach. This creates spaces in the thin-film, with the result that the refractive index of this porous thin-film is lower than that of the corresponding dense material.

Figure 2

A scanning electron micrograph of a silica film deposited by oblique angle evaporation reveals that feature sizes in the thin-film structure are in typically less than 100 nm (figure 2a, p28). This strongly limits the effect of Rayleigh scattering and so yields specular films of high optical quality.

What's really exciting about this technique is that the refractive index of the thin-film depends on the inclination of the nanorods, which in turn can be controlled by simply changing the deposition angle. Indeed, plotting the refractive index as a function of the deposition angle shows that the refractive index of silica can be tuned from 1.46 to 1.05 – very close to the refractive index of air – while for indium tin oxide (ITO) the refractive index can be varied from 2.1 down to 1.17 (figure 2b, p28).

This unprecedented ability to tune the refractive index over a wide range of desired values opens up a new paradigm in thin-film deposition. No longer is the refractive index of a material a fixed quantity. It can be controlled and tuned as required by the target application.

Particularly important is the technique's ability to attain extremely low refractive-index values, which offers the potential to realize completely new photonic structures. These include near-perfect anti-reflection (AR) coatings that are needed to get light into a material efficiently (in the case of solar cells) and out of a material (in the case of light-emitting diodes, LEDs). Furthermore, low refractive-index materials can be used to fabricate distributed Bragg reflectors (DBRs) from a single optical material.

Reducing reflections
A particularly promising application of these new optical materials lies in graded-refractive-index AR coatings. Conventional AR coatings consist of a single layer, a quarter-wavelength thick, that has a refractive index between that of the adjacent materials. However, it is well known that such AR coatings do not operate well over a broad range of wavelengths.

Figure 3

This deficiency can be overcome by fabricating AR coatings in which the refractive index is graduated from the low refractive-index material on one side to the high refractive-index material on the other. In the example shown in figure 3a, the coating consists of three layers of titanium dioxide (TiO₂)and two layers of silica deposited on an aluminium nitride substrate. The step-graded refractive index associated with the layers approximates the so-called modified quintic profile, which is well known to minimize normal-incidence Fresnel reflection.

Measuring the reflectivity of the coating at normal incidence for wavelengths ranging from 0.3 to 2.0 µm shows that the graded-index AR coating has low reflectivity over the entire visible and near-infrared spectrum (figure 3b). What's more, the reflectivity is less than 0.5% for wavelengths in the 574–1010 nm range. These results, which confirm the broadband nature of these coatings, could not have been attained with conventional quarter-wave coatings, or any other multilayer coatings.

Figure 3c compares the image of a computer monitor reflected by three samples: an uncoated piece of silicon; silicon coated with a quarter-wavelength coating of silicon nitride; and a graded-refractive-index AR coating made from TiO₂ and silica films. The darkness of the graded-index AR coating clearly demonstrates its superior anti-reflective properties over the conventional quarter-wave coating.

This broadband advantage is particularly important for photovoltaics applications, since the solar spectrum is inherently broadband – ranging from the ultraviolet (UV) to the infrared. Reducing the reflectivity over the full range of solar wavelengths increases the amount of light that reaches the active region, which in turn can have a significant impact on the energy-saving performance of photovoltaic cells. At the same time, incorporating graded-index AR coatings into LEDs would increase the amount of light coupled out of the active region and into free space, which could have a major impact on the light-extraction efficiency of these devices.

DBRs exploit single material
As well as near-perfect reflection-free coatings, low refractive-index materials can also be used for the opposite purpose: to make highly reflective mirrors, such as DBRs, made entirely of a single material, but with a high refractive-index contrast. It also means that materials for DBRs can be selected not just for their refractive index, but for other important material properties as well.

This offers a solution to the fact that the refractive index and other material properties, such as electrical conductivity and optical absorption, are in general inextricably coupled. And because only a limited number of optical materials are available, the choice of refractive index often dictates the other characteristics of the material.

For photonics applications that exploit both the optical and electrical properties of a material, this can force unsatisfactory compromises. In conventional DBR structures, the high- and low-index layers are composed of two separate materials, such as TiO₂ and silica. However, some materials do not work at certain wavelengths – TiO₂, for example, absorbs in the UV – while neither TiO₂ nor silica are conductive.

As a result, conductive DBRs are limited to epitaxially grown, doped semiconductors, which generally have low-index contrast. Conversely, DBRs composed entirely of ITO, which is both conductive and near-UV transparent, provide a good example of the design freedom afforded by oblique angle deposition.

Figure 4

This type of DBR can be fabricated by using oblique angle deposition to deposit alternate layers of low refractive-index ITO and normal ITO. In the example shown, three pairs of low-index/high-index layers have been deposited on a silicon substrate, with a scanning electron micrograph of the resulting DBR clearly showing the six individual layers (figure 4a, p30).

In this case the low-index layers consist of ITO deposited at an angle of 75°, while the highest refractive index is achieved for bulk ITO deposited at normal incidence. However, depositing bulk ITO at a near-normal incident angle on top of a highly porous nanorod film could cause some of the space between the nanorods to be filled. To avoid this potential problem, the high-index layers are deposited at an incident angle of –45°.

The micrograph reveals that the nanorod layers have feature sizes much smaller than a wavelength and also exhibit well-defined interfaces. As a result, the DBR is expected to have reflection characteristics that show good agreement with theory.

To put this to the test, the reflectance of two-pair and three-pair DBR structures made from ITO were measured as a function of wavelength (figure 4b). Also shown are the calculated theoretical reflectance spectra that use the complex refractive indices of the individual layers, which were measured on ITO films using ellipsometry.

It's clear that the agreement between theory and experiment is excellent. The measured real part of the refractive index is 1.49 for the low-index ITO and 1.90 for the high-index material. This means that the normalized index contrast (the difference in the two refractive indices divided by their average) of this ITO-based DBR is higher than in many semiconductor structures. The low-index ITO also has a low extinction coefficient – which allows light to pass through it with minimal absorption – while the extinction coefficient of the high-index ITO is close to that of the bulk material.

Commercial futures
To sum up, oblique vapour deposition allows us to fabricate a new class of optical thin-film materials with much lower refractive indices than would otherwise be possible. We have demonstrated a refractive-index value as low as 1.05 for a highly porous specular silica film, which also offers good optical quality.

These low refractive-index materials could offer serious benefits for a number of commercial applications. These include high-performance anti-reflective coatings for solar cells and LEDs, as well as a range of optical components, such as DBRs, that could benefit from being made of a single material. The next stage for the research will be to transfer the technology out of the research laboratory into the marketplace so that, someday, this new class of optical materials will become as common as conventional optical thin-film materials.

• This article originally appeared in the April 2008 issue of Optics & Laser Europe magazine.