Electron beam evaporation (without and with plasma assistance) as well as ion beam sputtering are used to prepare optical mixture coatings for applications in the ultraviolet spectral range. It is demonstrated that intermixing aluminum oxide/ aluminum fluoride materials by these physical vapor deposition techniques results in optical coatings with flexible refractive indices varying between 1.40 and 1.75 in the deep ultraviolet spectral region. At the same time, extinction coefficients vary between less than 1x10−4 and 2x10−3. For evaporated layers, at certain mixture ratios, mechanical stress appears to be close to zero.
© 2014 Optical Society of America
Aluminum Oxide (alumina, Al2O3) is one of the choices for mid-index coating materials typically applied in the wavelength range between 7000 and 200 nm . Moreover, because in the Vacuum Ultraviolet/Deep Ultraviolet (VUV/DUV) spectral regions, typical high index oxide materials such as titania, niobia, hafnia, tantala and the like show strong absorption and are therefore not suitable for interference coating design purposes, alumina becomes an interesting candidate for applications as a high index oxide VUV/DUV material . This is particularly relevant for applications at the 193 nm lithography wavelength .
While early data on the DUV optical constants of alumina films are published in , the principal status quo of alumina coatings produced by plasma ion assisted electron beam evaporation (PIAD) for 193 nm applications is well described in . Possible origins of remaining DUV absorption losses of high quality alumina coatings are discussed in [5,6].
In close affinity to alumina, aluminum oxifluoride has also reached the status of an optical coating material, however preferably produced by sputtering techniques [7,8]. Literature reports on practical applications of aluminum oxifluoride layers usually address antireflection applications in the IR or VIS [7–10], due to their higher flexibility in refractive index when compared to alumina.
Despite this, aluminum oxifluoride principally seems capable to act as a DUV/VUV material, when understanding it as a substance combining features of the DUV/VUV materials Al2O3 and AlF3 [11–13]. Moreover, both Al2O3 and AlF3 can be prepared by reactive Physical Vapor deposition (PVD) techniques, among them PIAD [14,15]. This offers the possibility to prepare oxifluoride films by fluorine-reactive evaporation (or fluorine-reactive PIAD) of alumina  as well as oxygen-reactive evaporation of aluminum fluoride .
An alternative approach to prepare mid- to low index UV layers with tailored refractive indices is offered by preparing mixtures of pure materials using PVD techniques. A combination of aluminum oxide and fluoride coating materials in mixture thin films is expected to offer an extension to lower refractive indices, higher optical band gap energies, and potentially higher laser induced damage threshold (LIDT) in the femtosecond regime compared to UV compatible oxide mixtures . Thus, Mende et.al. have reported successful preparation of ion beam sputtered (IBS) aluminum oxide/aluminum fluoride mixture coatings realized by co-sputtering from a zone target setup . In particular, an increase in the femtosecond LIDT of such single layer mixture films could be verified when comparing with pure alumina coatings.
In the present study, instead, emphasis will be placed on preparation of suchlike mixture coatings by co-evaporation in a PIAD process, while the focus is on the flexibility in optical (refractive index) and mechanical (layer stress) properties offered by these mixture coatings. Emphasis is further placed on reproducibility issues.
2.1 Film deposition
Al2O3/AlF3 mixture films have been prepared at Fraunhofer IOF in a Leybold Optics Syrus pro deposition system (equipped with an Advanced Plasma Source APSpro) by co-evaporation from commercial electron beam evaporators. AlF3 is evaporated from a 1–4 mm granulate supplied by Merck, while alumina is evaporated from 2.5 to 6 mm granulate supplied by Umicore. The mixture coating composition is controlled by the ratio of the corresponding evaporation rates, which are measured by means of quartz crystal microbalances.
In order to determine composition-dependent optical and mechanical film properties, 4 series of deposition experiments have been carried out (composition series). In each series, a set of mixture ratios has been realized by co-evaporation from two sources with a defined ratio of the corresponding deposition rates, as specified in Table 1. The specific details of each series are given by the type of working gas and level of assistance. Here, in the first series, samples have been produced by co-evaporation without any plasma assistance (Co-EBE), and without any reactive gas addition during deposition. In the second series, Co-EBE has been applied again, but now with addition of oxygen (oxygen flow 10 sccm). The third and fourth deposition series correspond to experiments with additional plasma assistance supplied by the APSpro (Co-PIAD) with a bias voltage of 100 V. Hereby, in the third series, assistance has been accomplished with an argon/oxygen mixture as working gas supplied through the plasma source, while in the fourth series, argon has been replaced by xenon (Ar-flow: 12 sccm, Xe-flow: 4 sccm). The target thickness was 300 nm and the substrate temperature 100°C in all deposition runs.
In each charge, one fused silica substrate (diameter 25 mm, thickness 1 mm) has been coated for optical measurements, either silicon substrates (25 mm x 25 mm, thickness 0.4 mm) or molybdenum substrates (diameter 10 mm, thickness 0.3 mm) for Energy Dispersive X-ray (EDX)-analyses and a silicon wafer (diameter 76.2 mm, thickness 0.4 mm) for stress measurement.
In order to place the obtained properties of Co-EBE and Co-PIAD prepared mixture samples into a proper context; we compare their optical properties with those prepared by means of IBS using a zone target setup (Co-IBS) at Laser Zentrum Hannover e.V. (LZH). The corresponding deposition procedure is described in . We note here, that Co-IBS has been performed with xenon as sputter gas in either an Ar/F2 or an oxygen-reactive atmosphere.
After having performed experiments as specified in Table 1 as well as suitable coating characterization, a second deposition experiment campaign at IOF was performed and focused on quantifying and improving the reproducibility. Therefore the optical and mechanical properties of a prospective type of mixture coating prepared by Co-PIAD with an argon/oxygen mixture were analyzed (reproducibility series). It is well known, that electron beam evaporation of alumina from a small crucible results in significantly improved deposition rate stability. For this reason the target physical thickness was reduced to 150 nm, while the thickness monitoring has been performed by the optical broadband monitoring system OptiMon  on witness glasses previously coated with a 250 nm thick tantala film for contrast enhancement. Optical monitoring has been performed assuming fixed optical constants as obtained from ex situ spectrophotometry; for samples which show a relevant vacuum-to-air shift, this may result in a systematic underestimation of the thickness measured in situ. This error has been accepted during this study, because the focus was here on reproducibility rather than on precision. Two series of deposition experiments have been performed to check the reproducibility of sample properties with a medium alumina filling factor palumina corresponding to approximately 62.5 vol.%. Each of the series consists of 5 subsequent deposition runs. The corresponding deposition conditions are presented in Table 2.
2.2 Film characterization
2.2.1 Energy dispersive X-ray-spectroscopy EDX of PIAD samples
EDX spectra have been recorded at IOF from all samples deposited in series 1-4 (Table 1) on either silicon or molybdenum substrates. EDX measurements have been performed using a high resolution scanning electron microscope FE-SEM Sigma (Zeiss). The applied silicon drift detector Inca X-act (Oxford Instruments GmbH) is specified with a resolution of 129 eV. Spectra have been analyzed using INCA Software (INCA Energy 250).
As the EDX-method is not particularly surface sensitive, the detection volume is dependent on the acceleration voltage and includes the film as well as parts of the substrate. In order to comply with the relevant substrate response, the film atomic composition has been estimated from a combined elaboration of two EDX spectra, obtained from different acceleration voltages (10 kV and 16 kV) for each sample. This way we hope to eliminate substrate contributions – at least partially.
EDX measurements of Co-PIAD samples have been performed with samples deposited onto silicon substrates. For the Co-EBE samples, molybdenum substrates have been available. In order to verify the compatibility of the results obtained from the two substrates, two of the Co-PIAD samples had been deposited on both types of substrates in the same deposition experiment; after having performed the correction procedure described below, the results in elementary composition differed from each other by less than 1.5 at% in both cases. Once the pursued variations in elementary composition as achieved by material mixing are much larger than these 1.5 at.%, we conclude that the comparison of the corrected results obtained from different substrates is reasonable.
The applied correction procedure works as follows: We start from (uncorrected) concentration raw data Nraw, element,voltage as obtained for the elements molybdenum Mo or silicon Si (substrate), aluminum Al, oxygen O, and fluorine F, which correspond to the full detection volume relevant for the acceleration voltage used. Thus, Nraw,Al,16 denotes the atomic concentration of aluminum as found from the 16 kV EDX spectrum. For each of the raw spectra (i.e. each voltage), we assume the normalization rule (here for Mo as substrate):
Then, corrected atomic concentrations Ncorrected, element valid for the film material are estimated from the subsequent application of following equations:
2.2.2 Optical properties
All samples have been characterized by means of UV/VIS spectrophotometry (near normal incidence transmittance and reflectance). From transmittance and reflectance spectra, refractive index n and extinction coefficient k as well as the film thickness d have been evaluated by standard curve fitting procedures based on a Lorentzian multioscillator approach . We emphasize that with respect to the Co-IBS samples prepared at LZH, applying the multioscillator approach is essentially a re-examination of the spectra previously published in , but this is necessary to guarantee comparability to the data originating from the Co-EBE and Co-PIAD deposition runs. Therefore, the n-data of Co-IBS samples published here may slightly differ from those published earlier in .
In order to quantify the shift (more precisely a combination of thermal and vacuum shift) of selected mixture coatings, we performed transmission measurements in the visible spectral range using the OptiMon process spectrophotometer .
These measurements have always been performed after more than one week of exposure to atmosphere, so that at least the large pores in the samples can be expected to be filled with water. First, a transmission measurement Tatm was performed in atmospheric conditions at room temperature. After that, the measurement chamber has been evacuated to high vacuum and heated up to a temperature of 100°C before making the second transmission measurement Tvac . From the differences in these spectra, the shift in optical thickness can principally be quantified. In most cases, this is done in terms of a relative wavelength shift of any prominent interference feature in the film spectrum, which is observed when the environment of the samples (temperature, humidity, pressure) is altered (see also ).
Unfortunately, this approach turns out to be ineffective when prominent interference features are absent because of a lack of a sufficient refractive index contrast. Because of the low refractive index and low thickness of the coatings mentioned here, instead of the shift as defined in [21,22], we make use of a modified but more general and practicable recipe of quantifying the changes in transmittance caused by heating and/or evacuating, which is given by Eq. (3) ( defines the wavenumber-equidistant grid of transmittance values):
For simplicity, we will still use the term “shift” for the value defined by Eq. (3), keeping in mind that it is now rather a shift in transmittance values along the ordinate axis than a shift along the abscissa axis in a transmission spectrum, as it is usually introduced in this context [21,22]. Nevertheless, in full agreement with the usual shift definition, according to Eq. (3), porous layers are again expected to show higher shift values than (almost) pore-free samples. In measurement practice, each of the transmission spectra Tatm and Tvac entering into Eq. (3) has been obtained after averaging over 50 spectral scans for noise reduction .
2.2.3 Stress measurements
For stress measurements, the curvature of uncoated silicon wafers has been determined by a Tencor system prior to deposition. After deposition of the film the measurement of the curvature has been repeated, and from the differences in curvature, the layer stress has been calculated by Stoney’s equation. In our convention, negative stress values correspond to tensile and positive ones to compressive stress.
3.1 Estimation of atomic concentrations by EDX
Figure 1 shows the results of the estimation of the elementary composition according to Sect. 2.2.1 of the samples specified in Table 1, as a function of the pursued alumina volume filling factor. On the right of this diagram (alumina filling factor = 1), no fluorine is present, so that these samples should correspond to a more or less stoichiometric aluminum oxide. In fact, the oxygen atomic concentration as found from EDX turned out to be somewhat higher than expected in stoichiometric alumina (60at.%). This extra oxygen amount is likely to be caused by some water which has penetrated into the pores of the film during storage at atmosphere, and which is not removed from the sample during the short duration of the EDX measurement. A decrease in the pursued alumina volume fraction is accompanied by a rapid increase in the fluorine atomic content, a clear decrease in the oxygen atomic content, and a moderate decrease in aluminum atomic content. All this is in qualitative consistency with what has to be expected. Nevertheless, even when no alumina is evaporated (Fig. 1 on the left), we still observe a residual oxygen signal of 5-7 atomic percent, which we assign again to a water contribution. Note that these 5-7 at.% are consistent with the extra oxygen amount found in the fluoride-free alumina coatings. A similar “residual” oxygen signal in aluminum fluoride has also been found by EDX for the IBS samples measured at LZH, as earlier reported in .
It is helpful visualizing the position of the data from Fig. 1 in a ternary compound diagram (Fig. 2). The red lines in Fig. 2 on the left indicate the recipe for how to read the diagram. Thus, the point P corresponds to a fictive mixture with 50 at.% fluorine, 25 at.% oxygen, and 25 at.% aluminum.
The experimental points corresponding to what is shown in Fig. 1 are presented in Fig. 2 on the right using the same color code as in Fig. 1. Here, all experimental points scatter in a region close to the theoretical prediction for a hypothetical mixture built from stoichiometric Al2O3 and AlF3 (full navy line). In the figure, corresponding theoretical dependencies are also shown for the hypothetical cases of mixtures of AlF3 with two different prominent aluminum hydroxides (boehmite, nordstrandite ), which may be present in any real alumina film as well. It seems nevertheless unlikely, that the experimentally observed behavior is dominated by the presence of the mentioned hydroxide modifications. The corresponding theoretical curves (in dash and dot in Fig. 2) have a descent different from what is observed in the experiment. The experimental behavior rather resembles the theoretical dependence of an alumina/aluminum fluoride mixture, with a constant offset which corresponds to some extra oxygen (at the cost of a reduction in aluminum content). This is consistent with the picture of a porous mixture layer, where the pores are filled with water, which leads to a decrease in the aluminum content accompanied by an extra oxygen fraction. Porosity has been verified independently by shift measurements (see Sect. 3.4).
3.2 Optical constants of the films
Keeping in mind that our primary purpose is the optimization of UV coating materials, we concentrate here on the optical constants obtained at a wavelength of 250 nm. In Fig. 3, these data are shown again as a function of the pursued alumina filling factor.
Figure 3 thus demonstrates that depending on the mixture ratio, the mixture coatings cover the refractive index range between the extremes of alumina and aluminum fluoride. The k - values of the mixture coatings show a more or less chaotic scatter around the “pure” alumina values, but it seems evident, that as a trend, Co-PIAD coatings show lower losses than Co-EBE coatings, while lowest extinction is achieved with Co-PIAD using Ar/O2 as working gas (the yellow squares).
In order to place these data into a proper context, it is useful to compare them to those obtained by Co-IBS, as well as other relevant data known from the literature. Such a comparison is shown in Fig. 4, where the determined k values are opposed to the corresponding n values in order to get a picture on the achievable combinations of UV optical constants. In this presentation, the available data appear segregated into two categories:
- Rather transparent layers, represented by the colored symbols. Co-EBE layers deposited with oxygen supply, as well as the Co-PIAD layers fall into this category. Here, we also find the Co-IBS prepared samples as well as data obtained from a study, where aluminum oxifluoride has been prepared by evaporation of alumina in an Ar-PIAD process with addition of fluorine as reactive gas . It is remarkable, that the IBS-data (using xenon as sputter gas) fall closest to the Xe-Co-PIAD data.
In Fig. 4, we recognize the expected trend that mixture coatings (no matter whether prepared by Co-IBS, Co-PIAD, or Co-EBE) appear principally capable to cover the full n- range between the indices of the pure materials which form the mixture. As it has been shown in , the same applies to the position of the UV absorption edge for mixture layers prepared by Co-IBS.
The oxygen reactive deposition of aluminum fluoride (grey triangles in Fig. 4) as well as the fluorine reactive deposition of alumina (green triangles in Fig. 4) also result in samples which cover a certain range of refractive indices; however, from available experimental data, we still observe a gap of UV refractive indices between 1.6 and 1.7 where no data points are observed.
The Co-EBE samples prepared without oxygen supply all appear strongly absorbing, which may be attributed to some lack of oxygen in the resulting films. We will return to this question later in the discussion. The more astonishing thing, however, is the strong absorption observed for the samples deposited by oxygen reactive evaporation or IAD of pure aluminum fluoride. Here, the data obtained from studies  and , which are certainly performed in different deposition systems, appear to be in an astonishingly good mutual consistency (grey triangles in Fig. 4), nevertheless the origin of the high absorption remains unclear. We plan to perform future experiments to clarify this point, while using exactly the same deposition system as utilized in the present study for Co-EBE and Co-PIAD.
3.3 Film stress
Film stress values as obtained for films from the series 1 – 4 deposited on silicon wafers are presented in Fig. 5.
We recognize that pure aluminum fluoride films all show a weak compressive stress, while evaporated pure alumina films show strong tensile stress. The weak plasma assistance as provided during deposition of series 3 and 4 results in an obvious reduction of the tensile stress. Mixtures result in coatings with intermediate stress values, i.e. some of the mixture films appear to be nearly free of mechanical stress.
3.4 Reproducibility considerations
Table 3 presents average values as well as the standard deviation for refractive index, shift (in the VIS), film thickness, and mechanical stress as obtained from 5 subsequent identical Co-PIAD deposition runs specified in Sect. 2.1 as reproducibility series.
Obviously, the stronger assistance (140 V) results in a lower shift than the corresponding 100 V series. Hence, the 100 V samples are expected to show higher porosity than the 140 V samples. This is confirmed by the higher refractive index and the compressive stress of the 140 V samples. We will now be able to estimate the influence of the bias voltage on the layer porosity, as shown in the discussion.
It further turned out, that for both values of the bias voltage; the shift data are nevertheless significant. In Table 3, we therefore observe the expected systematic discrepancy between in situ and ex situ determined film thickness values. This discrepancy becomes smaller when the bias voltage is increased, which is a typical behaviour.
First of all, we assign the established shift to some porosity, while the pores may either be empty (in vacuum) or filled with water (after storage at atmosphere). This way the value of the shift appears connected to some degree of porosity, which may be estimated when assuming a suitable mixing model for describing the optical properties of the coating.
There is much freedom in defining such a model, but the simplest approach is to assume a mixture of two components only: a pore fraction, and a solid fraction. The latter corresponds to a hypothetical pore-free mixture of alumina and aluminum fluoride. The packing density p will further be assigned to the volume fraction of that solid component. The porosity is then given by (1-p).
Different mixing formulas are available from the literature, corresponding to different assumed microstructures of the coating. Without detailed knowledge on the microstructure, a reasonable choice of the relevant model appears impossible. But for assumed real optical constants, the predictions of all relevant binary mixing models must fall in between the predictions of two extreme cases, known as Wieners bounds :
This way the knowledge about the shift behavior may be used to estimate upper and lower bounds of both layer porosity (1-p) and refractive index of the solid fraction nsolid. This is exemplified in the following for the Co-PIAD samples from the mentioned reproducibility series. Indeed, the refractive index of the 100V coatings is around 1.506 at 500 nm wavelength at atmosphere (Table 4). In order to theoretically reproduce the shift as experimentally determined in the VIS (Table 3), according to Eq. (3), we have to assume a refractive index in vacuum of 1.446. When substituting these n-values as nmixture into the first of Eqs. (4), and assuming further empty pores for the vacuum case (npore = 1.0) and water-filled pores for the case of samples exposed to atmosphere (npore = 1.33), a system of two equations is obtained that allows determining (1-p) and nsolid corresponding to the first of Wieners bounds. Repeating this procedure with the second of Eqs. (4), a second set of (1-p) and nsolid corresponding to the second of Wieners bounds is calculated. Assuming validity of these bounds to our system, the actual values of (1-p) and nsolid must be confined between the values obtained from the first and second of Wieners bounds. This way we obtain the following estimation of porosity and refractive index of the solid fraction (Table 4):
From Table 4 we recognize, that Wieners bounds result in a rather broad range of possible porosity levels that are still consistent with the observed shifting behavior. In particular, the porosity of the 100 V samples must be somewhere in between 8 and 23 vol%, while that of the 140 V samples is confined between 6 and 19 vol%. Note that these ranges are overlapping. The increase in the bias voltage from 100 V to 140 V is therefore not necessarily connected with a decrease in porosity (although this is a rather probable scenario), but in principle, modifications in morphology (shape or orientation of the pores) could explain the detected optical behavior as well.
We mention in this context, that the formal application of the Bragg Pippard mixing model to the refractive index data from Table 4 leads to a porosity of 12.4 vol.% for the 100V samples, and of 9.6 vol.% for the 140V samples. These data are well in between the bounds fixed by the application of Eqs. (4), and agree with reference data on the packing density of AlF3 and Al2O3 coatings published earlier in . Thus, porosity values reported for evaporated alumina films at 300°C substrate temperature account for up to 8vol.%; at lower substrate temperatures, a higher porosity is typically expected. Concerning aluminum fluoride, porosity data up to 38vol.% are reported (at 75°C substrate temperature ). The porosity values estimated from our samples (deposited at 100°C, but with some plasma assistance) fall in between the mentioned reference data and therefore appear consistent with earlier work.
The application of (4) to our experimental data also allows estimating the refractive index of the solid fraction of the films; corresponding to Table 4, we find a value of approximately 1.53 – 1.56 at 500 nm. This value is clearly in between tabulated values for densely packed AlF3 and Al2O3 coatings and appears thus reasonable.
In the present study, alumina/aluminum fluoride volume relations have been estimated from deposition rates recordings, while the elementary composition has been estimated from phenomenological corrected EDX-data. Both methods are expected to have limited accuracy, and in view of the considerable data scatter shown in Fig. 1, we do not expect a relative accuracy better than approximately 10% in both filling factor and elementary composition estimation. Therefore a discussion of the absorption losses in terms of possible stoichiometry variations appears to be difficult.
Nevertheless, on the basis of the available data, it is possible to create a diagram where at least a guess on stoichiometry-related absorption enhancement might be indicated. Indeed, let us return to Fig. 2 on the right. Here the EDX-results obtained from composition series 1-4 were collected, while the color code was used to distinguish samples by the deposition method. In this diagram, inaccuracies in deposition rate recordings do not matter. Let us now change the color code in order to distinguish the samples no more with respect to deposition method, but with respect to k @ 200 nm. The result is shown in Fig. 6.
In Fig. 6, we recognize the general trend that samples with lowest extinction (in black or red) tend to be arranged closer to the bottom of the ternary diagram than samples with highest extinction (in blue or violet). But that shift downwards is related to an increase in oxygen content at the cost of a decrease in aluminum content. Hence, as a very crude trend, Fig. 6 indicates that samples with lowest absorption have a higher O/Al ratio than samples with highest absorption. This trend is absolutely reasonable, because oxygen understoichiometry in any metal oxide or metal oxifluoride coating is well known to result in an absorption increase. On the other hand, all Co-EBE and Co-PIAD samples have been prepared in the same deposition chamber in otherwise identical conditions, so that we see no reason to attribute these differences in the DUV extinction to differences in the contamination level.
5. Conclusion and outlook
Mixture coatings as prepared by Co-EBE and Co-PIAD have been characterized mainly with respect to their DUV optical properties, and compared to corresponding results as obtained by Co-IBS. All of the mentioned deposition techniques allow preparing aluminum oxide/ aluminum fluoride mixture coatings (as verified by EDX) with flexible refractive indices varying between 1.40 and 1.75 in the deep ultraviolet spectral region. At the same time, extinction coefficients vary between less than 1x10−4 and 2x10−3. With respect to the n- and k-values, no remarkable difference can be obtained when comparing Co-PIAD and Co-IBS mixture coatings. As established in this study, some of the Co-EBE and Co-PIAD samples tend to appear practically free of mechanical stress because of certain porosity. Those mixtures appear prospective for use in DUV optical coatings with flexible refractive index and low mechanical stress.
A comparison of the data obtained in this study with those obtained from other sources leads us to the remarkable result that Co-PIAD and Co-IBS coatings tend to show lower extinction coefficients than oxifluoride coatings prepared by oxygen-reactive evaporation of aluminum fluoride. It remains unclear whether this is really a characteristic feature of the deposition technique used, or rather an accidental phenomenon resulting from different levels of contamination. We plan to reproduce the properties of such oxygen-reactively evaporated samples in a future work, by preparing them in the same deposition apparatus that has been used for Co-EBE and Co-PIAD experiments in the present study. This way we hope to generate new experimental material that is directly comparable to the results of the present study.
The authors are grateful to Heidi Haase for assistance in the EBE/PIAD deposition experiments at IOF. EDX measurements have been performed by Sylvelin Reinhard (IOF). Josephine Wolf (IOF) was kind enough to refine several of the figures. Financial support of the German ministries BMWi and BMBF in terms of the TAILOR (FKZ 16IN0665 and 16IN0667) and UltraLIFE (FKZ 13N11555 and 13N11558) grants is gratefully acknowledged.
References and links
1. M. Friz and F. Waibel, “Coating Materials,” in: Optical Interference Coatings, N. Kaiser and H. K. Pulker, Eds. (Springer-Verlag, 2003).
2. R. Thielsch, “Optical Coatings for the DUV/VUV,” in: Optical Interference Coatings, N. Kaiser and H. K. Pulker, Eds. (Springer-Verlag, 2003).
3. C. Zaczek, A. Pazidis, and H. Feldermann, in Optical Interference Coatings Topical Meeting, 2007 OSA Technical Digest Series (Optical Society of America 2007), FA1.
4. M. Zukic, D. G. Torr, J. F. Spann, and M. R. Torr, “Vacuum ultraviolet thin films. 1: Optical constants of BaF2, CaF2, LaF3, MgF2, Al2O3, HfO2, and SiO2 thin films,” Appl. Opt. 29(28), 4284–4292 (1990). [CrossRef] [PubMed]
6. J. Heber, C. Mühlig, W. Triebel, N. Danz, R. Thielsch, and N. Kaiser, “193 nm laser induced luminescence in oxide thin films,” Appl. Phys. (Berl.) 75(5), 637–640 (2002). [CrossRef]
7. G. L. Harding, “Production and properties of high rate sputtered low index transparent dielectric materials based on aluminium-oxy-fluoride,” Sol. Energy Mater. 12(3), 169–186 (1985). [CrossRef]
8. R. Lewin, R. P. Howson, and C. A. Bishop, “Optical coatings for large area interference filters,” Vacuum 37(3–4), 257–260 (1987). [CrossRef]
9. C. E. Anderson and J. P. Rousseau, Transparent Substrate Provided With, Thin-film Coating, U.S. Patent 5,952,084 (1999).
10. R. Korenstein, J. S. McCloy, P. E. Cremin, and R. W. Rustison, Durable antireflective multispectral infrared coatings, Patent 20100035036 (2010).
11. W. Heitmann, “Vacuum evaporated films of aluminum fluoride,” Thin Solid Films 5(1), 61–67 (1970). [CrossRef]
12. F. Bridou, M. Cuniot-Ponsard, J.-M. Desvignes, M. Richter, U. Kroth, and A. Gottwald, “Experimental determination of optical constants of MgF2 and AlF3 thin films in the vacuum ultra-violet wavelength region (60–124nm), and its application to optical designs,” Opt. Commun. 283(7), 1351–1358 (2010). [CrossRef]
14. O. Stenzel, D. Gäbler, S. Wilbrandt, N. Kaiser, H. Steffen, and A. Ohl, “Plasma ion assisted deposition of aluminium oxide and aluminium oxifluoride layers for applications in the ultraviolet spectral range,” Opt. Mater. 33(11), 1681–1687 (2011). [CrossRef]
15. M. Bischoff, O. Stenzel, K. Friedrich, S. Wilbrandt, D. Gäbler, S. Mewes, and N. Kaiser, “Plasma-assisted deposition of metal fluoride coatings and modeling the extinction coefficient of as-deposited single layers,” Appl. Opt. 50(9), C232–C238 (2011). [CrossRef] [PubMed]
16. X. Li, W. Zhang, J. Sun, J. Liu, Y. Hou, L. Lin, K. He, and K. Yi, “Influence of oxygen pressure and substrate temperature on the properties of aluminum fluoride thin films,” Appl. Surf. Sci. 282, 226–230 (2013). [CrossRef]
17. O. Stenzel, S. Wilbrandt, M. Schürmann, N. Kaiser, H. Ehlers, M. Mende, D. Ristau, S. Bruns, M. Vergöhl, M. Stolze, M. Held, H. Niederwald, T. Koch, W. Riggers, P. Burdack, G. Mark, R. Schäfer, S. Mewes, M. Bischoff, M. Arntzen, F. Eisenkrämer, M. Lappschies, S. Jakobs, S. Koch, B. Baumgarten, and A. Tünnermann, “Mixed oxide coatings for optics,” Appl. Opt. 50(9), C69–C74 (2011). [CrossRef] [PubMed]
18. M. Mende, I. Balasa, H. Ehlers, D. Ristau, D. B. Douti, L. Gallais, and M. Commandré, “Relation of optical properties and femtosecond laser damage resistance for Al2O3/AlF3 and Al2O3/SiO2 composite coatings,” Appl. Opt. 53(4), A383–A391 (2014). [CrossRef] [PubMed]
19. S. Wilbrandt, O. Stenzel, and N. Kaiser, “All-optical in-situ analysis of PIAD deposition processes,” Proc. SPIE 7101, 71010 (2008). [CrossRef]
20. O. Stenzel, S. Wilbrandt, K. Friedrich, and N. Kaiser, “„Realistische Modellierung der NIR/VIS/UV-optischen Konstanten dünner optischer Schichten im Rahmen des Oszillatormodells,” Vakuum in Forschung und Praxis 21(5), 15–23 (2009). [CrossRef]
21. O. Stenzel, S. Wilbrandt, N. Kaiser, M. Vinnichenko, F. Munnik, A. Kolitsch, A. Chuvilin, U. Kaiser, J. Ebert, S. Jakobs, A. Kaless, S. Wüthrich, O. Treichel, B. Wunderlich, M. Bitzer, and M. Grössl, “The correlation between mechanical stress, thermal shift and refractive index in HfO2, Nb2O5, Ta2O5 and SiO2 layers and its relation to the porosity,” Thin Solid Films 517(21), 6058–6068 (2009). [CrossRef]
22. P. W. Baumeister, Optical Coating Technology (SPIE Press Bellingham, 2004).
23. A. Petzold and J. Ulbricht, Tonerde und Tonerdewerkstoffe (VEB Deutscher Verlag für Grundstoffindustrie, Leipzig, 1983).
24. M. Held, Untersuchung maßgeschneiderter Nanokompositschichten für die Optik, PhD Thesis, in preparation.
25. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters, Springer Series in Material Science 25 (Springer-Verlag, 1995).