1.

GRATING MANUFACTURING USING MODERN MICRO STRUCTURING TECHNOLOGIES

Numerous processes can be used for grating manufacturing with different characteristics and grating profiles. Common processes for optical gratings are e-beam lithography [1], ruling, Direct-Laser-Writing (DWL) and holography. According to our experience holography or interference lithography is the best choice to generate almost arbitrary coherent grating structures on a manifold of substrate types. Further, on any substrate shape gratings with high peak efficiencies due to the option for blazing and least levels of ghost and stray light are feasible [5].

Holography is a parallel grating generation process that allows the pattern generation simultaneously on the entire optical surface [2-5]. As indicated in Fig. 1 a substrate may be curved as well. The handling of comparable heavy and/or very extended or thick or non-planar prism-like substrates is possible because there is in general no need to move the blank during the recording process. This is in contrast to the majority of the serial writing techniques. The common holographic approach is shown in Fig. 1, left side. This setup version generates generally sinusoidal symmetric grating profiles in the photoresist [6]. Especially if the depth of development is moderate – as indicated for example by the blue dotted curve in Fig. 3 left side – the resulting profile is close to an ideal sinus-shape. An optional subsequent ion etching process can be employed to transform these profiles into a blaze profile too. This approach has some drawbacks depending on the employed etching machine. In general, the angle of resulting blaze facets is globally oriented in the same direction. This leads to a more or less disturbing local variation of the grating efficiency. The situation gets worse if the surface angle variation of a curved substrate exceeds the magnitude of the blaze angle – as typical for EUV-gratings due to the comparable shallow blaze profiles.

Fig. 1:

The principle holographic setup approaches are shown. Left side: the symmetric holographic exposure setup, right side: in relation to the substrate surface the waves are primarily counter propagating

For the generation of blaze profiles directly in the resist Fig. 1. right side shows the adequate setup. This holographic approach is using counter-propagating waves in relation to the substrate surface. The geometries lead to the generation of a tilted stack of interference layers during the recording step. The final wet chemical development unveils a well approximated blazed profile [9, 10]. Distinguishing mark of these technique is a locally adapted blaze angle supporting the diffraction into a desired order in an optimal way [11].

The holographic principles, the resulting interference pattern and the typical photoresist profiles are shown in Fig. 1, Fig. 2, Fig. 3 and Fig. 4.

Fig. 2:

Left side: the interference pattern resulting from a symmetric holographic exposure setup, right side: interference pattern resulting from the opposing wave approach

Fig. 3:

Left side: photoresist grating simulation with a nearly symmetric profile due to holographic exposure setup in Fig. 1 left, right side: typical spatial modulation of solubility for counter propagating waves

Fig. 4:

Left side: deep developed photoresist grating SEM picture with a nearly symmetric profile (tilted symmetric exposure), right side: photoresist grating AFM picture (opposing wave approach)

2.

OPTIMIZATION INCLUDING CUSTOM SPECIFIC DEMANDS AND MANUFACTURING TECHNOLOGY

In a spectrometer systems optimization step it is beneficial if the optical designer doesn’t have to regard the current constraints of serial writing or ruling processes for the manufacturing of diffraction gratings. In this context holography provides a technological base for classical grating types like Offner- or Rowland-type gratings – even if the desired radii of curvature are comparable small. In general a variation of surface angles exceeding ±10 degrees within the clear aperture are manageable by holography as well. Further, concave imaging gratings or gratings defined by a variable line spaces definition (VLS gratings) are feasible. In general a close cooperation between grating manufacturer and the manufacturer of the spectrometer in early project states helps to find out the optimal grating type with regard to the systems performance as well as to a save technological path for the grating.

Proceeding from classical grating types, additional optical design parameters with various degrees of freedom can be used for the implementation of aberration corrected gratings and VLS (variable line space) gratings. Both types of gratings are highly appropriate for holographic exposure. Commonly a recording setup can be derived from any customers grating description based on substrate surface figure and line density distribution. Optional optimization loops incorporating all of the spectrometers optics may often lead to extra well adapted holography setups. This kind of a global optimization works with fixed parameters for the elements of the spectrometer. Further, proceeding with the initial merit function the line density distribution will be still close to the classical grating description. Additionally, the potential for bending the lines and/or use higher order coefficients in case of a VLS grating can be employed for achieving enhanced focusing properties or imaging quality.

Beside the optimization of the imaging performance, the parameters efficiency and polarization sensitivity will influence the system performance of spectral remote sensing systems [7]. The manufacturing method of the grating also predefines the type of the grating profile. Therefore the recording type has to be set regarding the desired spectral energy efficiency. All mentioned grating types above are available with symmetrical grating structures or with blazed profiles depending on what is the best choice for the application. A tuning of the spectral efficiency curve can be achieved by scaling the initial profiles in the height dimension.

The often preferred initial (resist-) blazed profile lead in most cases to more complex recording setups including expensive special optical parts. During a conception phase ray tracing based simulations are used to simulate and optimize the holographic recording setup with its individual optical elements including the calculation of tolerances. The design of the exposure optics is adapted to achieve highest imaging quality with respect to blaze profile, correct grating line orientation and groove density variation. Beside this optimization, based on the customer’s spectrometer design an angular range of interest (field of view) can be defined in order to optimize the holographic setup towards unwanted back-reflections or false light, which can lead to unwanted satellites with minor intensity [7]. In most cases, the optimization of the holographic setup can be realized by using special designed optical elements for compensation of aberrations while avoiding these unwanted back-reflections. In order to secure the optical performance of this often preferred blaze profile even on complex grating types an adapted inspection technique has to be provided.

3.

WAVEFRONT MEASUREMENT RESULTS OF AN ABERRATION CORRECTED GRATING

Plane and curved gratings with moderate curvature and tiny deviations from an equidistant line distribution can be tested by a conventional interferometer. The typical interferometer measurement for optical surfaces employs the 0th order of a sample. To assess the optical performance of a grating or a holographic element one of the accessible higher diffraction orders is captured. This approach is limited to wave front curvatures that lead to line densities below the Nyquist-frequency of the measuring system – mainly limited by the pixel sizes of the detector array. Thus, in practice many grating types - including concave imaging gratings, Offner-gratings or aberration corrected and VLS-gratings - are not measurable by this method. Therefore we use adapted strategies to ensure the quality of the huge variability of diffractive elements manufactured by our in house technology.

Fig. 5:

left side: a typical Moiré-pattern generated by a test setup for an Offner-grating in an even interference field, right side: extended pattern to achieve a square matrix that is beneficial for the numerical processing – the numbers on the x- and y- axes indicate pixel numbers

Fig. 6:

Moiré lines transferred into a surface figure of a corresponding wave front, by this method an optical designer is able to predict the performance of a spectrometer system – the information about the wave aberration is coded in the distance and curvature of the fringes – the examples shows a PV-value for the wave aberration of about 0.8 periods (note the edges have to be neglected)

More challenging test objects are gratings with or without curvature and with a comparable strong variation of the line density along the grating vector. Though it is in some cases a possible way to include a CGH in the optical path of an interferometer VLS as well as aberration corrected gratings in general offer another more flexible way to test them accurately. Here we show an example based on a precisely aligned static interference field (according to the Offner-grating measuring setup described above) combined with a VLS grating sample. Fig. 7 a) shows the Moiré-pattern for this blazed VLS grating on a toroidal substrate embedded in the projected clear aperture. The grating design exhibits a symmetrical line distribution in grating line direction at its vertex. It was manufactured with a holographic recording process and due to the additionally holographic aberration correction, the sample has neither equidistant nor straight grating lines. Therefore an even periodic test field leads only to a locally constrained phase match area indicated by the elliptical zero-interference region (zero-Moiré-region) close to the center of the figure.

Fig. 7:

a) Moiré-pattern of the whole clear aperture of the VLS grating (circular, diameter 100 mm);

b) Magnified region centered to the best matching area between test field and local grating line density with indication of meaning;

c) Moiré-pattern with grating rotation of 0.1000 degrees in relation to Fig. 5 a)

d) Derived line density distribution – the variation on the sample grating is about 0.5%

The initial alignment of the interference field is careful carried out by using a reference grating. The spatial frequency of the interference field has to be at least as high as the highest line density on the grating surface. After this alignment procedure one will observe at least a small grating region that fits to the spatial frequency of the test field as indicated by the examples in Fig. 7. Then a defined change of the period of the interference field relative to the sample surface generates a local shift of the best matching area. A rotation axis perpendicular to the grating vector in the vertex leads to a lateral shift along the dispersion direction of the grating. The local matching between interference field and grating frequency can be modified preferably by a defined rotation of the grating (Fig. 7 c) or a rotation of a plane wave in the interference field or of both waves. If the rotation continues the best matching area will migrate further until a set of sampling points - sufficient for a retrieval of the line density distribution - is gathered. Detecting the location of this zero-interference-region precisely and associate it to the corresponding rotation angle an accurate assignment of grating line density and orientation is feasible. In Fig. 7 b) the distance “a” corresponds to a shift up to a (phase) mismatch of 0.5 grating lines whereas “b” indicates the mismatch of a whole period of the grating. The center detection of the zero-Moiré-region can be supported by suitable fit functions in an evaluation step. Thus a sub-pixel accuracy can be achieved and this in turn corresponds to a very small fraction of the grating period (or wave length under measuring conditions). By this method we get a sensitivity comparable to a conventional interferometer only by detecting the centers of the best matching areas. Unlike to the interferometer measurement this approach needs a final evaluation step to retrieve the wave aberration from a set of discrete sampling points. However, despite the necessity of graphic rectification of the captured images due to the variation of projection conditions caused by moved/rotated optical components in the setup, it is based on today’s typical computer performance a comparable fast and easy way.

4.

SUMMARY

The present paper reports about two different approaches of holographic setups – to create symmetrical and blazed resist profiles - and their adaption for the generation of tailored grating structures even on curved substrates. The prerequisite of an adapted measurement technique to complete the holographic process chain is a central statement. Our new approach in its basic configuration is able to identify the wave front aberration of a classical Offner type grating. An additional option for an adaption of the interference field to locally varying line densities enables the investigation of VLS gratings as well. The measurement technique works virtually independent of the substrate shape and its accuracy is in the range of lambda/20 using the existing test setup. The methods became a standard measurement technique developed at ZEISS during the past years. They can be used for the prediction and analysis of the spectrometer performance employing the investigated gratings.