Optical Design of a High Flux Setup in the Extreme Ultraviolet

. Abstract. Extreme ultraviolet (XUV) light applications are still a very promising field that was heavily enlivened by the definition of the new wavelength for semiconductor lithography within the XUV range. But the detection of XUV light is also important for the exploration in the field of space science (i.e., monitoring the formation and evolution of solar storms) and high-energy physics (i.e., dark matter detection). The advancement of this technology mainly depends on the performance optimization of XUV sources, optical systems and related photodetectors. In this work, the optical design of a high flux XUV setup was simulated and defined to optimise the beam path which is the backbone of the initial evaluation process for the characterisation of luminescent materials under XUV irradiation. Additionally, the paper focused on the conceptualisation and realisation of the experimental setup as well as the alignment of the optical components and the detector calibration.


Introduction
Many techniques are currently utilized for developing compact table-top radiation sources to provide highly coherent extreme ultraviolet (XUV) and soft X-ray pulses that can open new areas of applications and a deep understanding of light featurization.The aim of this work is to design and realise a tabletop tool and define the related process.The tool should as well allow for measurements with high intensities in the order of those expected in the mask plane in high-volume manufacturing EUV lithography (~1 W/cm²) [1,2], also this would enable measurements of the saturation and degradation of luminophores under practical conditions.

Methodology
To measure the absolute quantum efficiency of luminophores under XUV excitation, the tool must allow to illuminate the sample with a known amount of XUV photons and measure the output of luminescence light from the sample with a known collection efficiency.For the degradation measurements, a light source with high power output and corresponding focusing optics is required.For such high-intensity measurements, it is also required to control the illumination time with good precision to avoid unwanted overexposures of the sample.Additionally, a vacuum environment is needed to minimise the absorption of the XUV radiation.The tool requirements can thus be defined as follows; 1-light source for XUV radiation, 2-spectral filtering for EUV radiation in case of a broadband light source, 3-monitoring of the light amount impinging on sample, 4-detection of luminescence light from sample, 5-optics for light collection and focusing on samples, 6-precise control of illumination time, 7-vacuum environment.

Optical Design and Realisation of the Setup
The largest influence on the layout of the tool has the light source and the optics as they define the beam path that all other components must be tailored around.If the optic is designed specifically for the setup it can be of course adapted to other critical components.The pulsed xenon plasma discharge source is used as light source.These sources produce a plasma pinch with a submillimetre diameter and strong emission lines in the XUV [3].As these sources have a broad emission spectrum up into the infrared, therefore spectral filtering is required.Two-shell ruthenium-coated Wolter collector is used to collect and focus the light from XUV source, as shown in Figure 1.The collector is used to demagnify the source into the sample plane that can maximise the achievable intensity.The design values for the geometry of the collector are listed in Table 1.To find the optimal distance of the source and sample relative to the collector, the plasma source and the collector are simulated within the Zemax OpticStudio 15, as shown in Figure 1(b,c).The software allows to optimise the position of source and sample plane as a minimisation of a given merit function.The spatial distribution of the light ray origins of the source is simulated as a Lorentzian distribution along and perpendicular to the optical axis.The simulation results in a spot on the sample with 50 μm FWHM, as illustrated in Figure 2, which corresponds to a demagnification of the image of the plasma pinch by a factor of seven.Together with the known xenon spectrum () and frequency  of the source, its light output within a 2 % bandwidth around 13.5 nm and the total spectral filtering ML()⋅Zr(), the effective light output of the setup (Ω) and expected energy per pulse on the signal diode (Ediode), calculated as shown in Figure 4.
The result of such a calibration measurement is shown exemplarily in Figure 5 for a measurement without collector and an illumination aperture of 5 mm diameter in a distance of 1908 mm to the source (that corresponds to 105 mm behind the simulated focal plane).Figure 5 (a) shows an average charge per pulse of around 300 pC for the signal diode and 630 pC for the reference diode.This difference is expected as the reference diode is positioned much closer to the source.The responsivity of the diode for EUV light is given by the manufacturer: ℛ(13.5 nm) = 0.27 A/W.In general, it would be more precise to integrate the responsivity times the spectral intensity of the source over the wavelength but the difference here is minimal and the manufacturer reports an uncertainty of 10 % on the responsivity anyway.The ratio of the photodiode charges is shown in Figure 5 (b).As expected, the variations of the diode charges shown in Figure 5 (a) are highly correlated.Nonetheless, the scattering of the values around the mean was not anticipated.While the reference diode looks onto the source under an angle of 4.3° relative to the optical axis, one would only expect an overall relative difference to the amount of light emitted along the optical axis.

Concluding Remarks
Calibration measurement shows such a relative difference would just lead to a different ratio of the diode signals that are then incorporated into the calibration factor.This behaviour may be explained by the movement of the plasma pinch between pulses.Depending on their position, the diodes then receive more or less light with no constant ratios.While the mean ratio is still a reproducible value, especially for longer measurements (> 5 min), this behaviour may require additional attention.

Fig. 1 :
Fig. 1 : (a) Scheme of a type 1 Wolter telescope, (b) expected focus point and (c) simulation of the used two-shell Wolter collector.

Figure 2 :
Figure 2: Simulation of the light spot from the collector in the plane of highest intensity (a) at 50 μm and (b) at 100 μm behind that plane.The absolute intensities shown in Figure2are calculated based on the light output of the source and the absorption/reflectivity of the additional optical components.The simulation predicts a peak intensity of 0.85 W/cm².The FWHM of the spots are (a) 50 μm and (b) 140 μm respectively.The total power in both spots is around 33 μW.The hexagonal shape seen in (b) is a result of the light blocked by the six collector holder bars.The collector also allows for a so-called flat top distribution of the light, illuminating a larger area with roughly constant intensity.Figure3shows a photograph and a scheme of the assembled setup.

Figure 4 :
Figure 4: Effective light output of the setup (Ω) and expected energy per pulse on the signal diode (Ediode) at different source frequency (f).

Figure 5 :
Figure 5: Calibration measurement of the signal diode and the reference diode measured with an integration unit.Each data point is the average over 10 pulses.

Table 1 :
Design values for the two-shell Wolter collector.