Open Access
Add to BibSonomyAbstract
In this work, the influence of using ion beam sputtered mixtures instead of pure materials and the impact of applied post deposition annealing to residual stress is investigated. Single layer pure films and mixtures of Nb2O5 / SiO2 as well as multilayer coatings are examined by the means of residual stress. High residual compressive stress was measured for all as-deposited samples. Pure and mixed monolayer samples were annealed at various temperatures and residual stress was determined after each annealing routine. Residual changes in optical constants, layer thickness and surface roughness upon annealing are examined to explain stress behavior. Obtained data was used to make optimization of high reflectivity structures with completely eliminated residual stress. The proposed method can be used to coat very thin substrates where flatness requirements are essential.
© 2016 Optical Society of America
1. Introduction
Laser beam quality is an essential characteristic in many scientific and technological fields which can be defined in several ways as well as determined by several factors. From the point of view of optical elements pulse wavefront distortion is strongly dependent on optical surface quality which is usually affected by applied optical coatings. Wavefront error can be compensated by adaptive methods or phase masks but these methods are expensive and time consuming so alternative solutions for film induced distortion must be implemented. Thin film induced stress is fundamental problem related to optical coatings, leading to deterioration of both optical, mechanical characteristics and performance of optical systems. Thin film stress depends on many factors including deposition technique, process parameters, substrate and post deposition treatment. For ion beam sputtered coatings high compressive stress is typical [1] while other standard deposition methods can lead to either tensile or compressive stress. Induced stress in thin films may lead to delaminating, fracture or microstructural changes [2,3]. Besides total film destruction stress may induce variation in optical coating performance: shift / change transmittance / reflectance spectrum [4,5] or curve the substrate shape. It has been shown that stress can be manipulated by several methods. All these methods can be divided in two categories: in–situ, meaning that the method is applied during coating process and ex–situ, meaning that method is used after deposition process. Materials having tensile and compressive stress can be used together in combination to reduce stress in multilayer system [6]. Nevertheless, the choice for such combinations is limited even for standard technologies. Deposition parameters like temperature, speed, ion assistance and partial gas pressure can be adjusted to influence microstructural film formation and thus the residual stress. It has been also shown, that mixture material coatings experience lower internal stress compared to bulk ones [7,8]. Mixing several materials during deposition can also be a solution to reduce stress even for IBS (ion beam sputtering) induced stress [1]. All the above mentioned methods can be called in situ. Additional back side coatings can compensate stress for most multilayer structures [9]. Being economically ineffective method it is not used very often, especially for complex shape optical elements. Argon ion machining [10] and film cracking [11] methods were used for stress compensation, although they are destructive methods and are not applicable in all cases where structure of coating cannot be changed. Thermal annealing as ex-situ process is one of the simplest and most widely used [12–14]. It is prevalent due to minor changes of the coating optical and mechanical properties (even improvement of optical properties in some cases may be achieved) [12]. Although, this method has limitations too. For different materials appropriate annealing temperatures and rates must be determined empirically. In addition, annealing to high temperatures can lead to delaminating or fracture caused by mismatch in thermal expansion coefficients between substrate and coating, microstructural changes can increase loss, induced by scattering mechanisms. In this study a combination of two above mentioned common stress compensation techniques leading to simple, versatile and efficient procedure is demonstrated. Complete stress compensation is demonstrated for NIR high reflectors made by Nb2O5 / SiO2 based mixed materials. Detailed results for single layer coatings are examined and discussed.
2. Experiment
Samples for experiment were prepared using ion beam sputtering machine from Cutting Edge Coatings (Fig. 1.). A radio-frequency argon ion source was used to strike metal target, consisting of two different materials at an angle of incidence of 57°. The target was mounted on linear translation stage at a distance of 25 cm from substrates, thus enabling shifting ion-target interaction zone across two different coating materials: Si and Nb. Mixtures were made by pointing the ion source in to the intersection between them. Ion acceleration voltage was up to 1400 V for all materials and mixtures, resulting in deposition speeds of 0.85 Å/s for Nb2O5 and 1.3 Å/s for pure SiO2. During process, 50 sccm oxygen gas was fed near the substrates to ensure complete oxidation of the film and thus resulting the working pressure of ~3*10−3 Pa. Sputtering process was performed in a temperature of ~50 °C.
Fig. 1 Process chamber (~1 m3 volume): 1 – drive for substrate rotation, 2 – shutter, 3 – primary ion source, 4 – substrates palette, 5, 6 – target and translation stage, 7 – assist ion source, 8 – gate valve, 9 – high vacuum pump, 10,11 -materials, 12 – joint edge.
Broadband optical monitoring was used to control physical thickness during deposition process and thus samples of 4*λ/4@1064 nm optical thickness were produced. Transmission spectra in range of 200-1100 nm were recorded with spectrophotometer Perkin Elmer Lambda 950 after deposition. Monolayer coating refractive indices and volumetric fractions were determined from recorded transmission data using OptiChar software [15]. For refractive index determination low absorption spectral zones were investigated and monolayers were considered as low or nonabsorbing. Volumetric fraction of selected materials in mixture layers were calculated according to Bruggemann’s formula [16] and using index dispersion data of selected pure monolayer coatings. Thermal annealing was performed in SNOL 8.2/1100 heating furnace ensuring temperature repeatability of less than 1°C. Sputtered films were annealed in air to 200 °C, 300 °C, 400 °C, and 500 °C temperatures. Optimized multilayer mirrors were annealed at predefined temperatures in range of 300-450 °C. In all cases single annealing routine consisted of three parts: heating phase, constant annealing phase and cooling phase. During heating phase temperature was linearly raised by 100 °C per hour until determined annealing temperature was reached. At constant annealing phase temperature remained stable for one hour. During cooling phase temperature was reduced by 100 °C per hour until room temperature was reached. Substrate curvature was measured after each annealing routine. For flatness and stress determination of monolayer samples Veeco Dektak 150 profilometer was used. Thin film mechanical stress was calculated from substrate curvature data using Stoney formula [17]. 1 mm thick and 25,4 mm in diameter fused silica substrates with initial flatness (substrate flatness was determined in Peak to Valley (PV) values) of PV<λ/6, where λ = 633 nm, was used. In both the stress and the surface flatness measurements 24 mm aperture was investigated. Tensile stress was considered as negative and compressive stress was considered as positive in all cases. Flatness of substrates with multilayer coatings before and after annealing was measured by Zygo Verifire XPZ interferometer. Analysis of interferometric fringes was performed by Zygo Mx software. AFM measurements were performed using Veeco Dimension Edge device and tapping mode for all sample to indicate surface roughness changes upon annealing.
3. Results and discussion
8 different in volumetric fraction single layer coatings of Nb2O5 / SiO2 were produced. Recorded transmission spectra depicted in Fig. 2. Optical constants of pure Nb2O5, SiO2 and mixtures were defined by fitting transmission spectra using Cauchy and exponential models for refractive index and extinction coefficient respectively. Both models as well as fitting procedure are implemented in OptiLayer software package [15]. Volumetric fraction in mixtures was defined by fitting transmission spectra with a Bruggemann’s model - changing fraction of pure materials to get best fit. Least squares fit was used in all cases. Good agreement between measured transmission and model was observed while defining optical constants and fractions, which suggest non-essential errors in fraction and dispersion determination. Common features of mixed oxide coatings like blue shift of absorption edge in UV range [18] and scaling of refractive indices according to volumetric fraction [8], are evident and depicted in Fig. 3.
Fig. 2 Transmission spectra of fused silica substrate and single-layer mixtures of Nb2O5 and SiO2 according to volumetric fraction of Nb2O5.
Fig. 3 Extinction coefficient (a) and refractive index dispersion (b) of Nb2O5 and SiO2 mixtures according to volumetric fraction of Nb2O5.
After annealing a change in thickness and optical constants was established. An increase in physical thickness by 1.1-1.4% was observed for all single layer samples after annealing to 300 C°. Meanwhile, a decrease in average refractive index by 0.4-0.6% was also determined from transmittance spectra. No changes in absorption or scattering were detected using spectral measurements after annealing up to 500 °C. No clear trend with material fraction was observed for above mentioned changes. Resulting increase of approximately 1% in optical thickness is in good agreement with other published papers and can be explained with a void formation under high temperature annealing of IBS coatings [19,20]. Void formation was also confirmed by a slight increase in surface roughness according to AFM measurements. Profilometric measurements of substrate’s surface bending were performed on fused silica samples of 1 mm thickness, designated for stress determination. To reduce measurement error and impact of initial surface irregularities (PV < λ/6), measurements were performed in several directions and only averaged values were used in further calculations. Residual film stress was calculated using Stoney equation:
where ts is substrate’s and tf is film’s thickness respectively, R2, R1 – radius of curvature before and after deposition respectively, E = 72 × 103 N/mm2 and ν = 0.17 - Young’s modulus and Poisson coefficient of the fused silica substrate respectively [21]. Calculated residual stress for Nb2O5 / SiO2 mixtures before and after each annealing step is depicted in Fig. 4(a). Margin of error up to 20 MPa can be expected as a result of surface irregularities and measurement accuracy. It is notable, from these results, that pure SiO2 has larger impact on both resulting average stress in multilayer and curvature of the substrate as thickness of SiO2 layers are larger than Nb2O5. It also reduces its amplitude by a factor of 2 only after annealing at approximately 500 °C. On the other side, pure Nb2O5 shows much smaller compressive stress which is almost equal to SiO2 after annealing at 500 °C. Nevertheless, after annealing Nb2O5 changes the sign of stress at 300 °C and leads to high tensile stress. In this case stress amplitude after annealing to 500 °C seems to be saturated and is almost equal (but the sign is opposite) comparing to initial - as deposited state. For mixed materials resulting stress can be made even smaller than for pure ones. This result was also observed in before mentioned publications [7,8] and is related to microstructural differences between pure and mixed materials. Several stress components can be distinguished as a drop in compressive stress for SiO2 and Nb2O5 is slightly different. While thermal expansion coefficients of FS (0.55 × 10−6/C°) and SiO2 should be similar a drop in stress after annealing can be related to relaxation of compressed structure which is in agreement with increased thickness. In the case of Nb2O5 which has larger CTE (5.8 × 10−6/C°) a slightly larger reduction of stress is observed. This difference in stress behavior can be assigned to the thermal stress induced by a different CTE of substrate and coating. It is also well known that incorporation of small fraction of other material prevents thin oxide film structure from crystallization effect [8] which was not observed at any of samples mentioned in the paper (Fig. 4(b).).Fig. 4 Residual stress according to Nb2O5 fraction in monolayer mixture at different annealing temperatures. AFM measurements of pure Nb2O5 before and after annealing to 500 °C (b).
A modeling of resulting stress in multilayer coatings was made based on results in Fig. 4(a). An assumption was made that stress in multilayer is an average of stresses in single layers of the stack and is calculated by formula [22]:
where σM is total stress in multilayer coating, σH,L is stress and tH,L is total thickness of high (H) and low (L) refractive index materials in mirror. Several possible mixture’s combinations were selected to make highly reflecting mirror with a theoretical reflectivity R larger than 99.9% for 1064 nm wavelength at normal angle of incidence. Pure or mixed materials for mirror were selected according to these establishments: minimization of required annealing temperature and keeping theoretical reflectance R > 99.9%. In this study, as one would expect, mixed materials proved to serve the purpose better than the pure ones, however, leading to increased total number of layers as well as total thickness of mirror. According to results in Fig. 4, as high refractive index material (H) 93% mixture was used. On the other side, as a low index material (L), several choices, leading to different annealing temperatures or total thickness exist. The lower fraction of Nb2O5 in L material would lead to higher annealing temperature but less layers to reach specified reflectivity. The coating design with 93% and 8% of Nb2O5 has been chosen as optimal, so that residual stress would become zero or negligible after annealing to 430°C. Theoretical prediction, using data from Fig. 4., of resulting stress for multilayer coating with a design structure (HL)^14L using pure materials and optimized mixtures according to annealing temperature for fused silica substrates is plotted in Fig. 5. A clear stress decrease with increased annealing temperature is present. Linear extrapolation of stress above 500 °C for HR coating based on pure materials is presented in dotted line. For a complete stress compensation of the structure a temperature of approximately 575°C is required. At this temperature range effects like crystallization, which would increase scattering loss of the coating can be expected [8,20,23]. Selected optimized multilayer reflector coating was sputtered on 1 mm thickness fused silica sample with initial PV<λ/6 and on 2 mm sample of PV<λ/10 flatness but same material to ascertain better understanding of method limitations. During the process both substrates were located at the same radius of rotation in the tooling, thus ensuring the same thickness of the coating. The same post deposition annealing procedure as for single layer coatings was performed for mirrors. Resulting residual stress is plotted in Fig. 5. A small discrepancy between predicted and measured stress may arise from errors in initial determination of stress for single layer coatings and zone target position calibration error leading to change in mixtures volumetric fraction. However, all these factors lead to the change of annealing temperature, required to completely compensate surface bending by only 10 °C.Surface flatness interferometric measurements of aforementioned samples are depicted in Fig. 6. Measurements before and after annealing at 430 °C are only analyzed. Multilayer coatings on FS samples before temperature treatment exhibit high compressive stress resulting in measured PV values close to approximately 3λ and 10λ for 2 mm and 1 mm thickness samples respectively. However, after annealing at 430 °C, compressive stress in SiO2 dominated low index layers is overcompensated by tensile stress in Nb2O5 dominated high index layers, and thus resulting in PV values lower than λ/10 for 2 mm substrate (Fig. 6.) and equal to uncoated surface irregularities for 1 mm thick substrate.
Fig. 6 Surface mapping of as deposited mirror (a) and after post deposition annealing (b) on 2 mm fused silica sample.
4. Conclusion
It has been shown that stress in single mixed layers depend on material fraction and annealing temperature. Changes in residual stress after annealing were explained by differences in CTE (between substrate ant thin monolayer) and layer material structural changes resulting in modified optical constants and increase of surface roughness. In this study we proved that IBS multilayer coatings can be made with completely compensated residual stress using a combination of thermal annealing and material mixing. It has been shown that compressive stress in SiO2 rich mixture layers is compensated by tensile stress in Nb2O5 rich mixture layers after annealing at a certain temperature without any noticeable changes in both losses and material structure. Thus, it is possible to make stress compensated coatings without second surface coatings or accurate and time consuming sputtering process parameter optimization. In some cases even initial substrate surface spherical deviations can be compensated using accurate control of post deposition process. In further research, the method demonstrated here, can be applied to other material combinations and deposition technologies or other substrate materials with different thermal expansion and mechanical coefficients. Application to some specific thin film optical problem solution would be appropriate for demonstration.
Acknowledgments
We would like to thank Optical coating laboratory at CPST for permission to use sputtering and thin film characterization equipment.
References and links
1. C. J. Tang, C. C. Jaing, K. S. Lee, and C. C. Lee, “Residual stress in Ta2O5-SiO2 composite thin-film rugate filters prepared by radio frequency ion-beam sputtering,” Appl. Opt. 47(13), C167–C171 (2008). [CrossRef] [PubMed]
2. L. B. Freund and S. Suresh, Thin Film Materials: Stress, Defect Formation and Surface Evolution (Cambridge University Press, 2004).
3. G. N. Strauss, “Mechanical Stress in Optical Coatings,” in Optical Interference Coatings, N. Kaiser, and H. Pulker, eds. (Springer Berlin Heidelberg, 2003).
4. S. Shigeki, “Temperature dependence of transmission characteristics of multilayer film narrow bandpass filters,” Jpn. J. Appl. Phys. 38(1), 6362–6368 (1999). [CrossRef]
5. T. C. Chen, C. J. Chu, C. H. Ho, C. C. Wu, and C. C. Lee, “Determination of stress-optical and thermal-optical coefficients of Nb2O5 thin film material,” J. Appl. Phys. 101(4), 043513 (2007). [CrossRef]
6. J. B. Oliver, P. Kupinski, A. L. Rigatti, A. W. Schmid, J. C. Lambropoulos, S. Papernov, A. Kozlov, C. Smith, and R. D. Hand, “Stress compensation in hafnia/silica optical coatings by inclusion of alumina layers,” Opt. Express 20(15), 16596–16610 (2012). [CrossRef]
7. B. J. Pond, J. I. Debar, C. K. Carniglia, and T. Raj, “Stress reduction in ion beam sputtered mixed oxide films,” Appl. Opt. 28(14), 2800–2805 (1989). [CrossRef] [PubMed]
8. A. Melninkaitis, T. Tolenis, L. Mažulė, J. Mirauskas, V. Sirutkaitis, B. Mangote, X. Fu, M. Zerrad, L. Gallais, M. Commandré, S. Kičas, and R. Drazdys, “Characterization of zirconia- and niobia-silica mixture coatings produced by ion-beam sputtering,” Appl. Opt. 50(9), C188–C196 (2011). [CrossRef] [PubMed]
9. T. Amotchkina, M. K. Trubetskov, Y. Pervak, L. Veisz, and V. Pervak, “Stress compensation with antireflection coatings for ultrafast laser applications: from theory to practice,” Opt. Express 22(24), 30387–30393 (2014). [CrossRef] [PubMed]
10. T. G. Bifano, H. T. Johnson, P. Bierden, and R. K. Mali, “Elimination of stress-induced curvature in thin-film structures,” J. Microelectromech. Syst. 11(5), 592–597 (2002). [CrossRef]
11. M. Ohring, The Materials Science of Thin Films (Academic Press, 1992).
12. M. Bischoff, T. Nowitzki, O. Voß, S. Wilbrandt, and O. Stenzel, “Postdeposition treatment of IBS coatings for UV applications with optimized thin-film stress properties,” Appl. Opt. 53(4), A212–A220 (2014). [CrossRef] [PubMed]
13. A. Hallbauer, D. Huber, G. N. Strauss, S. Schlichtherle, A. Kunz, and H. K. Pulker, “Overview about the optical properties and mechanical stress of different dielectric thin films produced by reactive-low-voltage-ion-plating,” Thin Solid Films 516(14), 4587–4592 (2008). [CrossRef]
14. R. Thielsch, A. Gatto, and N. Kaiser, “Mechanical stress and thermal-elastic properties of oxide coatings for use in the deep-ultraviolet spectral region,” Appl. Opt. 41(16), 3211–3217 (2002). [CrossRef] [PubMed]
15. “OptiLayer optical coating software,” http://www.optilayer.com/.
16. T. Tolenis, M. Gaspariūnas, M. Lelis, A. Plukis, R. Buzelis, and A. Melninkaitis, “Assessment of effective-medium theories of ion-beam sputtered Nb2O5–SiO2 and ZrO2–SiO2 mixtures,” Lith. J. Phys. 54(2), 99–105 (2014). [CrossRef]
17. G. Stoney, “The Tension of Metallic Films deposited by Electrolysis,” Proc. R. Soc. Lond., A Contain. Pap. Math. Phys. Character 82(553), 172–175 (1909). [CrossRef]
18. M. Lappschies, M. Jupé, and D. Ristau, “Extension of Ion Beam Sputtered Oxide Mixtures into the UV Spectral Range,” in Optical Interference Coatings, OSA Technical Digest (Optical Society of America, Tucson, Arizona, 2007), paper. TuA7.
19. J. T. Brown, “Center wavelength shift dependence on substrate coefficient of thermal expansion for optical thin-film interference filters deposited by ion-beam sputtering,” Appl. Opt. 43(23), 4506–4511 (2004). [CrossRef] [PubMed]
20. E. Cetinörgü-Goldenberg, J. E. Klemberg-Sapieha, and L. Martinu, “Effect of postdeposition annealing on the structure, composition, and the mechanical and optical characteristics of niobium and tantalum oxide films,” Appl. Opt. 51(27), 6498–6507 (2012). [CrossRef] [PubMed]
21. M. J. Weber, Handbook of Optical Materials (Taylor & Francis, 2002).
22. A. E. Ennos, “Stresses Developed in Optical Film Coatings,” Appl. Opt. 5(1), 51–61 (1966). [CrossRef] [PubMed]
23. V. Janicki, J. Sancho-Parramon, S. Yulin, M. Flemming, and A. Chuvilin, “Optical and structural properties of Nb2O5–SiO2 mixtures in thin films,” Surf. Coat. Tech. 206(17), 3650–3657 (2012). [CrossRef]
