1.

INTRODUCTION

In recent years telescopes based on near normal-incidence multilayer mirror technology have been employed in many missions dedicated to the Sun observation in EUV wavelengths, as in particular Fe-IX (17.1 nm), Fe-XII (19.5 nm), Fe-XV (28.4 nm) and He-II (30.4 nm). Examples of successful missions are SOHO (EIT) [1] and TRACE [2]. Performance of multilayer are mainly evaluated in terms of peak reflectivity at working wavelength and rejection capability of unwanted lines.

Because of their good time stability, Mo/Si multilayer are conventionally used for all of these wavelengths, even if, at the longer ones they have relatively low reflectivity peak with respect to other material couples. Moreover, in the case of a Mo/Si periodic structure optimized for 28.4 nm, the reflectivity curve is quite spectrally broad and includes the strong HeII line, that can affect diagnostic with the Fe line signal. A suitable narrowband solution which cuts down the HeII reflection has been obtained through an a-periodic ML structure design [3]. Moreover, further enhancement of peak reflectivity can be obtained using other material combinations like Mg/SiC [4], B₄C/Si/Mo [5] or B₄C/Si [4]. Although for some of these structures, test and proof of lifetime stability must be investigated.

We propose an innovative method for designing suitable capping layer covering the multilayer structure which do not affect the reflectivity peak while rejecting unwanted emission, for example from relatively close lines. The capping layer solution can be adopted both in case of periodic and a-periodic multilayer, made by different materials. The capping layer can be realized using a ML structure of different materials, for example to get additional mechanical or optical properties, as the capability of surviving in harsh environmental space conditions or to suitably reject visible/UV spectral ranges.

In this paper we present and discuss theoretical results for some structures designed by the use of the new mathematical tool. Experimental results related to periodic Mo/Si multilayer covered by an optimized Mo/Si capping layer able to reflect the Fe-XV line with rejection ratio of some orders of magnitude for the near He-II intense line are presented.

In section 2 a theoretical analysis of the capping layer design is exposed, in section 3 some theoretical results are discussed. In section 4 preliminary experimental results are presented.

2.

ANALYTICAL DESIGN METHOD

The innovative basic idea is to take advantage of the e.m. field standing wave configuration generated in the multilayer structure by the superposition of incident and reflected fields. As already pointed out in [6], the last protective layer in a ML structure can be grown at the node-position of the standing wave field intensity distribution in the ML. In fact, by suitable design of the last uppermost layers it is possible to shift the standing wave distribution at the top of the ML. In this way the performance of the structure result essentially insensitive to the cap-layer characteristics.

Let’s now consider two wavelengths, the first λ_peak is the “useful” wavelength and the second λ_noise is the wavelength to be rejected. ML (multilayer) is the coating sub-structure constituted by the repetition of two or more materials designed in order to obtain the best reflectivity peak at the λ_peak wavelength and CL (capping layer) is the structure made of last layers covering the ML, which is designed in order to preserve the λ_peak wavelength signal and suppress the λ_noise wavelength signal (see Fig. 1).

Fig. 1

A schematic view of the conceptual subdivision between CL and ML and of their performance.

The CL structure, like the ML structure, is constituted by a sequence of absorber and spacer materials (see Fig. 1), the materials can be the same or different from the ones in the ML, in any case their thickness must be suitably optimized.

The optimization design sequence of an optimal structure consists of the following steps (see Fig. 2):

Fig. 2

A schematic of the sequence of steps followed for the multilayer optimization, here, for clarity, only the standing wave at λ_peak is shown. Step 1, ML design; step 2, Fields computation; step 3, CL design; step 4, optimization.

By Positioning the absorber layers into the standing wave node we minimize the detrimental radiation extinction effect at the λ_peak wavelength almost preserving the same reflectivity peak of the ML structure. At the same time, due to the different standing wave behavior at different wavelengths, we can have an higher radiation extinction at the λ_noise wavelength than at the λ_peak one.

3.

SIMULATION OF SOME APPLICATIONS

Different structures have been designed for reflecting the 33.5 nm or the 28.4 nm lines while rejecting the strong 30.4 nm one. These structures are based on Mo/Si ML structures with different CL, and are reported in Table 1. In this section we show and discuss the theoretical simulations, performed with IMD program [7]. The considered cases can be very interesting for example for GOES-R mission [8].

Table 1:

column 1, the index of the different cases, column 2 and 3, respectively the ML and CL materials, in column 4 and 5, respectively the λpeak and λnoise wavelengths in nm.

3.1

Case 1

Mo/Si ML with W/Si CL working at 33.5 nm with high rejection at 30.4 nm have been designed. The W absorption coefficient is reported in Fig. 3, it shows a very high extinction at both wavelengths of interest, in addition thin W layer deposition has been already tested for ML structures for X-ray mirrors.

Fig. 3

The behavior of the W extinction coefficient in the 20-45 nm spectral range.

In Table 2 the structure of the optimized multilayer is reported. In Table 3 and in Fig. 4 the performance of the optimized structure is compared with the performance of a standard Mo/Si periodic multilayer. The new design shows a peak reflectivity loss of 2.6% in absolute percentage with respect to the standard periodic ML but with a considerably improved, about two orders of magnitude, rejection ratio.

Table 2

The structure of the optimized multilayer for case 1.

CL Structure	Value
W	2.0 nm
a-Si	16.5 nm
W	2.2 nm
a-Si	16.5 nm
W	2.0 nm
ML structure	Value
Period (a-Si/Mo)	18.2 nm
Ratio	0.89
Period number	35

Table 3

Columns 2 and 3, respectively the reflectivity at the λpeak and λnoise wavelengths for the case 1.

	R33.5 nm	R30.4 nm
Standard periodic	0.197	0.044
Optimized ML+W/Si CL	0.171	1.79*10-4

Fig. 4

In the case a) The reflectivity behavior of the optimized multilayer with a W/Si CL, continuous curve, compared with the performance, dashed line, of a standard periodic multilayer optimized only for the λ_peak wavelength. In the case b) the same data of the case a) reported in log scale. The reflectivity are optimized and calculated at 5° normal incidence.

3.2

Case 2

The next three cases are based on Mo/Si ML structures with different CL, designed in order to reflect at 28.4 nm with the highest rejection at 30.4 nm.

In this case a Pt/Si CL has been designed. Pt has been chosen because it is a very suitable CL absorber material for this spectral region (see Fig. 5).

Fig. 5

Pt extinction coefficient in the 25-35 nm spectral range.

The structure is reported in Table 4, in Table 5 and in Fig. 6 the performance of the optimized multilayer are compared with a Mo/Si periodic multilayer designed in order to reflect the 28.4 nm wavelength. In this case we have obtained a reflectivity loss of 4% in absolute percentage but with improved rejection of about three orders of magnitude.

Table 4.

The structure of the optimized multilayer for case 2.

CL Structure	Value
Pt	2.2 nm
a-Si	13.4 nm
Pt	2.0 nm
a-Si	12.8 nm
Pt	2.0 nm
a-Si	11.6 nm
Mo	2.0 nm
ML structure	Value
Period (a-Si/Mo)	15.3 nm
Ratio	0.87
Period number	35

Table 5

Columns 2 and 3, respectively the reflectivity at the λpeak and λnoise wavelengths for cases 2,3,4.

	R28.4 nm	R30.4 nm
Standard periodic	0.25	0.045
Optimized ML + Pt/Si CL	0.21	6.7*10-5
Optimized ML + Cr/Si CL	0.2	3*10-5
Optimized ML + Mo/Si CL	0.196	2.62*10-5

Fig. 6

In the case a) The reflectivity behavior of the optimized multilayer with a Pt/Si CL, continuous curve, compared, with a standard periodic multilayer optimized for the λ_peak wavelength, dashed line. In the case b) the same data of the case a) reported in log scale. The reflectivity is optimized and calculated at 5° normal incidence.

3.3

Case 3

The next two cases have been chosen in order to have an easier deposition procedure, i.e. with a lower number of sputtering target materials.

In this case a Cr/Si CL has been adopted. A Cr adhesion layer between the substrate and the multilayer can be grown in order to avoid any adhesion failures, moreover, the Cr layer tends to absorb surface contaminants on the substrate, and it also has tensile stress which balances the large compressive stresses in the multilayer. The high stress in the multilayer is due to high gamma ratio (γ) value of the Si/Mo coatings optimized for this spectral region.

In this case the Cr material has been used both like CL absorber layer and adhesion layer, this in order to permit the deposition of an optimized multilayer structure by utilizing a magnetron sputtering deposition system with only three cathodes.

Cr doesn’t show the best property as candidate for the CL structure, in particular the absorption is lower than for the Pt or Mo cases. For this reason in the optimization process the Cr layers are slightly displaced with respect to the 28.4 nm standing wave nodes. The new positions give the best ratio value between the standing wave area into the CL structure at the 28.4 nm and 30.4 nm wavelengths.

The resulting structure is reported in Table 6, we highlight the high number of CL layers necessary to obtain a high enough rejection, due to the low absorption of Cr. The resulting performance is reported in Table 5 and in Fig. 7

Table 6

The structure of the optimized multilayer for case 3.

CL Structure	Value
Cr	3.2 nm
a-Si	12.3 nm
Cr	3.05 nm
a-Si	12.6 nm
Cr	2.9 nm
a-Si	12.8 nm
Cr	2.7 nm
a-Si	13.1 nm
Cr	2.5 nm
a-Si	12.1 nm
Mo	2.0 nm
ML structure	Value
Period (a-Si/Mo)	15.2 nm
Ratio	0.87
Period number	35

Fig. 7

In the case a) The reflectivity behavior of the optimized multilayer with a Cr/Si CL, continuous curve, compared with, a standard periodic multilayer optimized for the λ_peak wavelength, in dashed line. In the case b) the same data of the case a) reported in log scale. The reflectivity are optimized and calculated at 5° normal incidence.

3.4

Case 4

In this last case only Mo and Si have been taken into account as possible materials both for the ML and CL structures.

Mo has a relevant absorption coefficient in this spectral region, it assures a good time stability coupled with Si, only one Mo layer for the CL structure placed in an optimal position that corresponds to a standing wave node for the 28.4 nm wavelength and standing wave anti-node for the 30.4 nm wavelength has been chosen. The multilayer structure is shown in Table 7, in Table 5 and in Fig. 8 the optimized multilayer performance is compared with the standard periodic multilayer.

Table 7

The structure of the optimized multilayer for case 4.

CL Structure	Value
a-Si	14.7 nm
Mo	2.2 nm
a-Si	57.75 nm
Mo	2 nm
ML structure	Value
Period (a-Si/Mo)	15.15 nm
Ratio	0.868
Period number	35

Fig. 8

In the case a) The reflectivity behavior of the optimized multilayer with a Mo/Si CL, continuous curve, compared with the performance, of a standard periodic multilayer optimized for the λ_peak wavelength, dashed line. In the case b) the same data of the case a) reported in log scale. The reflectivity are optimized and calculated at 5° normal incidence.

4.

EXPERIMENTAL RESULTS

In this section preliminary experimental results are presented. Samples deposition has been performed at RXO with magnetron sputtering technique. Preliminary tests have shown that samples with structure as for case 3, Mo/Si ML + Cr/Si CL, are quite critical. In fact the resulting performance can be very sensitive to manufacturing tolerances, due to the relatively high number of Cr layers; even a small error in the various layer thickness can result in a not negligible displacement of the Cr layers with respect to the standing wave field distribution. Furthermore preliminary experimental tests have shown that the ML+CL structure performance is also critically dependent on the optical constants of materials, accordingly an optimized structure design has been derived using the optical constants experimentally measured and reported by Tarrio et al. [9]. The corresponding Mo/Si multilayer with Mo/Si CL structure is reported in Table 8. A prototype of this sample has been deposited (RXO) and tested. The reflectivity has been measured with a laser plasma facility [10].

Table 8

The structure of the re-designed multilayer using optical constants of Tarrio et al.[9], see text.

In Fig. 9 the experimental results for the a-periodic optimized structure, dotted curve, compared with the theoretical simulation, dashed curve, and those computed for a standard periodic structure, continuous curve, are reported.

Fig. 9

Experimental results for the a-periodic optimized structure (dotted curve), compared with the theoretical simulation (dashed curve), and those computed for a standard periodic structure (continuous curve).

The good agreement between theoretical and experimental results is noticeable. Simulations show that the critical parameter of this multilayer is the thickness of the second 41.2 nm thick Si layer, a relative error of only a few percent can affect considerably the final performance, resulting in lower reflectivity peak and shifting of peak and rejected wavelengths. From these preliminary measurements the good agreement between the theoretical and experimental reflectivity wavelength peak confirms the high accuracy of this critical layer.

Further measurements are planned at ALS synchrotron in order to accurately measure the rejection ratio at the 30.4 nm wavelength, the measurements here reported are, in fact, limited by the signal to noise ratio.

5.

CONCLUSIONS

We have presented an innovative method for the design of multilayer structures with improved spectral filtering performance. Structures with high rejection at the strong 30.4nm HeII line have been designed. Preliminary samples have been realized and tested, obtaining good agreement with simulations and demonstrating actual feasibility.