Open Access
Add to BibSonomy
Abstract
Atomic layer deposition (ALD) has been proven as an excellent method for depositing high-quality optical coatings due to its outstanding film quality and precise process control. Unfortunately, batch ALD requires time-consuming purge steps, which leads to low deposition rates and highly time-intensive processes for complex multilayer coatings. Recently, rotary ALD has been proposed for optical applications. In this, to the best of our knowledge, novel concept, each process step takes place in a separate part of the reactor divided by pressure and nitrogen curtains. To be coated, substrates are rotated through these zones. During each rotation, an ALD cycle is completed, and the deposition rate depends primarily on the rotation speed. In this work, the performance of a novel rotary ALD coating tool for optical applications is investigated and characterized with ${{\rm SiO}_2}$ and ${{\rm Ta}_2}{{\rm O}_5}$ layers. Low absorption levels of ${\lt}{3.1}\;{\rm ppm}$ and ${\lt}{6.0}\;{\rm ppm}$ are demonstrated at 1064 nm for around 186.2 nm thick single layers of ${{\rm Ta}_2}{{\rm O}_5}$ and 1032 nm ${{\rm SiO}_2}$, respectively. Growth rates up to 0.18 nm/s on fused silica substrates were achieved. Furthermore, excellent non-uniformity is also demonstrated, with values reaching as low as $\pm {0.53}\%$ and $\pm {1.07}\%$ over an area of ${135\times 60}\;{\rm mm}$ for ${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$, respectively.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. INTRODUCTION
Atomic layer deposition (ALD) is a well-established and high-quality thin-film deposition technique with several unique features. Among these features, the self-limitation growth mechanism gives rise to excellent thickness control and dense films. These enables many applications, for example, in the semiconductor industry [1–3], photonics [4,5], and optoelectronics [6]. Batch ALD is the most common implementation of the ALD principle. A vacuum chamber is sequentially filled with precursor A that reacts with the surface of the heated substrate. The precursor is chemically designed in such way that it reacts only with the surface groups present on the substrate. It reacts neither with the atmosphere inside the vacuum chamber nor with atom species on the surface that have already reacted with a precursor molecule, ultimately leading to a self-limited growth [7]. Once all the surface sites on the substrates have reacted with the precursor A, all the unreacted precursor molecules and reaction products are purged away with a flow of inert gas. Depending on the ALD process variant used, a plasma or second precursor B is used to react with the precursor A saturated surface, forming a chemical surface function which is able to react again with the first precursor A. Finally, a second purging step is used to remove reaction educts and products. With a repetition of this cycle, thicker layers are formed, as shown in Fig. 1.
As a result of the self-limited growth, dense films can be produced with an extraordinary control of the film thickness. Furthermore, the gaseous nature of the precursors leads to pinhole-free and highly conformal growth even on complex or structured surfaces, giving this technique its unique properties [8]. Besides common applications in semiconductor production, ALD-grown films have shown low optical losses and high laser damage thresholds [9], which make this technique highly interesting for the production of optical coatings. Thus far, low growth rates have restricted ALD to thin antireflective coatings, which make process development and production time-intensive [9,10]. In the course of the rapid development of photonics and laser technology, an ever-increasing demand for strongly curved lenses, high-power laser systems, structured optics, and environmentally stable materials have led to a rising interest for ALD coatings. To satisfy the need for a faster ALD system, the field of spatial ALD has been developed. Herein, the time-sequential process of batch ALD is transferred to separate spatial areas inside the ALD reactor.
In this work, we present high-quality and low-loss ${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$ optical coatings deposited with an ultra-fast rotary spatial ALD process. We investigate the compositions, optical properties, and film uniformities of these materials and show that this process exhibits high potential as an alternative technology for fabricating optical coatings in industrial-scale applications.
2. EXPERIMENTAL SETUP
Figure 2(a) illustrates the implementation principle of rALD. While the principle seems simple, the realization of this process is significantly more complex than it may appear. Figure 2(b) shows a picture of the chamber configuration with optical substrates fixed on a rotating table.
Fig. 2. Working principles of rALD. (a) Schematic drawing of reaction zones inside the rALD reactor. (b) Photograph of the rALD reactor from the top.
The growth rate of the films mostly depends on the movement of the substrates through each of these zones. Clearly separated reaction zones are needed to avoid a reaction of precursors A and B directly with each other, which would lead to undesired side effects such as particle formation and chemical vapor-like deposition [11,12].
Four sources of precursor A are available in this setup, whereas three reaction gases are available for the plasma source. The tool was specified to coat a substrate height up to 30 mm. 200 mm silicon wafers were used to study wafer non-uniformity, whereas 25 mm quartz glass samples were used for optical analysis. For the experiments presented here, (tert-butylimido)tris(ethylmethylamido)tantalum (TBTEMT) grade 4 N from EpiValence Ltd. and bis(diethylamino)silane (SAM.24) electronic grade from Air Liquide Electronics GmbH were used for ${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$, respectively. Reaction gas ${{\rm O}_2}$ with grade 5.1 was deployed as oxidation agent and 750 mA plasma current was applied. The reaction chamber was heated to 150 °C and the rotation speed was varied between 100 and 200 rpm. Finally, nitrogen was used as the purging gas.
To measure the uniformity of the process, the spectral reflection of the ALD-coated 200 mm wafers was measured with a self-built spectrometer system integrated on a $xy$-translational stage table. The spectral reflection measurements and calculations of layer thicknesses were based on the software of the in-house developed Broad Band Monitoring (BBM) System [12]. The non-uniformity (NU) is given by
where ${d_{{\max}}}$ is the maximum and ${d_{{\min}}}$ the minimum thickness of the measured area. The average of minimum and maximum thickness is denoted ${d_{\rm{average}}}$.Photospectral analysis was performed in transmission mode with a PerkinElmer LAMBDA 1050 UV/Vis/NIR spectrophotometer at the wavelength range of 200–2000 nm to determine the refractive index and layer thickness. Sellmeier coefficients were modeled with a least-squares algorithm. Modeling was done with absorption. The absorption of the coatings was determined experimentally by laser calorimetry at a wavelength of 1064 nm according to ISO 11551:2019 [13]. For online monitoring of the film growth, our in-house developed BBM system [12] was integrated in the C2R-machine and connected to the machine control. This system allows observation of the coating progress and execution of a precise endpoint detection for film thickness in a multilayer coating. In this way, complex and thick film designs can be deposited precisely. This system is also resilient against rate fluctuations, which might be caused by variations in precursor supply, due to a low precursor level in the supply bottle for instance. This is important, because precise monitoring of the filling level of precursor bottles is not possible during the coating process. Without an online monitoring system, coating runs with an unnoticed empty precursor bottle can waste expensive substrates and precursor material of the second container used for the layer system and potentially cause long supply times. In contrast to online ellipsometry measurements, the transmission-based approach is tolerant of low height variations, making this system ideal for this rotating spatial ALD approach.
Fig. 3. BBM integration. (a) Schematic image of BBM setup (beam propagation in green). (b) Online detected grown film thickness.
Figure 3(a) illustrates the working principle of BBM: A broadband light source is collimated and led on a path through the ALD machine into a spectrometer. Once every rotation, the light is transmitted through the coated optic and is affected by the spectral thin-film characteristics. The measured spectrum is compared to a dark measurement and a measurement without an optical sample. A thin-film design with expected dispersion data is compared to the measured data and the current thickness is calculated as shown in Fig. 3(b). Furthermore, refractive index and thickness can be calculated online with the BBM software modules. An advantage of the rotary ALD approach in contrast to thermal ALD is that coatings are only deposited on the rotary table and substrates, which transits both reaction zones, while in thermal ALD all surfaces with suitable reaction or decomposition temperature can be coated including cover glasses. In the C2R tool, the light path is located in the purging area, where no coating occurs. The system actually has a spectral range form 400–1000 nm but can be extended to 240–1700 nm.
3. DEPOSITION AND RESULTS
To investigate if rALD is usable for optical applications, film morphology, composition, uniformity, spectral performance, and absorption of ${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$ thin films was measured and is discussed in this section.
A. Film Morphology and Composition of${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$
As shown in Figs. 4(a) and 4(c), a specific surface morphology could not be resolved with a magnification of 15000 and higher. Particles in the pictures are shown to ensure the correct focus on the surface and were hard to find. These particles were likely introduced during storing and handling of the coated substrates before mounting them into the SEM. Cracks, impurities, or holes could not been found in any sample. In addition, there were no signs of crystallization.
Furthermore, to get an impression of the elemental composition of the films, an energy-dispersive x-ray (EDX) analysis was performed. The main signal peaks of both films, which are shown in Figs. 4(b) and 4(d), indicate silicon and oxygen. This can be explained with the quartz substrate and, in Fig. 4(b), also by the ${{\rm SiO}_2}$ coating. In Fig. 4(b), signals of tantalum (Ta) are also present, referring to the coated ${{\rm Ta}_2}{{\rm O}_5}$. In both spectra, the signal of carbon can also be found. This is typical for SEM investigations, because the electron beam is cracking hydrocarbons present in the SEM chamber as part of the rest atmosphere. The result is a thin carbon film on the sample surface. Nevertheless, it should be noted that the carbon signal of the tantalum pentoxide sample is higher than the one on the silicon dioxide sample. This can indicate a part of the carbon resulting from difference between the coatings. One explanation could be that the ligand of TBTEMT contains more carbon, which is incorporated into the coating in low quantity. Overall, the influence of carbon is low, demonstrated by the low absorption values for ${{\rm Ta}_2}{{\rm O}_5}$ and also for ${{\rm SiO}_2}$.
No additional signal peak could be found, indicating no further materials of relevant concentration are contained in the films.
B. Film Uniformity
To gain insights into the uniformity of rALD coatings, two $8^{\prime \prime}$ silicon wafers were coated with around 200 nm of coating materials ${{\rm SiO}_2}$ and ${{\rm Ta}_2}{{\rm O}_5}$ (Fig. 5). Both wafers were coated with 100 rpm at 150 °C reactor temperature. After coating, the spectral reflection was measured in a raster with a spacing of 4 mm between each point on the $xy$ table. From the reflection measurement, the coated single layer was modeled and its calculated thickness is shown in Figs. 5(a) and 5(b). Coordinate point ($x = {0}\;{\rm mm}$, $y = {100}\;{\rm mm}$) is the closest point to the rotation axis. Measurement points are marked with black dots and the color in between the dots is an interpolation. While the highest thickness for the ${{\rm SiO}_2}$ coating can be found close to the table rotation axis, the highest thickness value of the ${{\rm Ta}_2}{{\rm O}_5}$ coating occurs in the outer region of the substrate and therefore at a higher radius to the table rotation axis.
Fig. 4. SEM and EDX analysis of ${{\rm Ta}_2}{{\rm O}_5}$ and ${{\rm SiO}_2}$ samples. (a) SEM image of ${{\rm Ta}_2}{{\rm O}_5}$ coating on polished quartz glass. (b) EDX analysis of 200 nm ${{\rm Ta}_2}{{\rm O}_5}$ coated on quartz glass. (c) SEM image of ${{\rm SiO}_2}$ coating on quartz glass. (d) EDX analysis of 200 nm ${{\rm SiO}_2}$ coated on quartz glass.
Fig. 5. 2D thickness distribution illustration on a 200 mm silicon wafer deposited at 150 °C with 100 rpm for (a) ${{\rm SiO}_2}$ and (b) ${{\rm Ta}_2}{{\rm O}_5}$.
Fig. 6. Photospectral measurements and modeled layers of rALD coatings. Black, measurement; green, modeled layers. (a) 186.2 nm ${\rm Ta_2O_5}$ coating, (b) 193.6 nm ${\rm SiO_2}$ on a ${\rm Ta_2O_5}$ pre-coated quartz substrate, and (c) 1032 nm thick ${\rm SiO_2}$ fast coating on a ${\rm Ta_2O_5}$ pre-coated quartz substrate uncoated quartz substrate (magenta).
For the ${{\rm SiO}_2}$ coating, a smaller circle inside Fig. 5(a) gives a better impression of uniformity. For a diameter of 140 mm, a thickness deviation of 12.61 nm, corresponding to a non-uniformity of $\pm 3.49\%$, was found. For the rectangular shaped area of ${135\times 60}\;{\rm mm}$, the deviation was determined to be 3.94 nm, resulting in a non-uniformity of $\pm {1.07}\%$.
In contrast to this, the ${{\rm Ta}_2}{{\rm O}_5}$ coating in the same areas show a lower non-uniformity for the 140 mm circle in Fig. 5(b) of 3.64 nm, corresponding to a non-uniformity of $\pm {0.86}\%$, while the rectangular area of ${135\times 60}\;{\rm mm}$ has a deviation ranging around 2.25 nm, corresponding to $\pm 0.53\%$.
For the whole 200 mm diameter wafer, a non-uniformity of $\pm {4.76}\%$ for ${{\rm SiO}_2}$ and $\pm {1.64}\%$ for ${{\rm Ta}_2}{{\rm O}_5}$ was measured. Pfeiffer et al. have shown $\pm {1.5}\%$ for ${{\rm SiO}_2}$ and $\pm {4.0}\%$ for ${{\rm Ta}_2}{{\rm O}_5}$ coatings on the same-sized substrates for batch ALD [8].
The precursor and temperature distributions are a main source of non-uniformities in ALD processes. In rALD, temperature distributions only become an issue if they are a function of the radius, as the rotation homogenizes distribution on one radius. Nevertheless, precursor distributions appear as a more urgent problem due to the short interaction time during a rotation cycle. The difference between the two materials might be due to differing mobilities of the precursor’s molecular size. Nevertheless, rALD reaches nearly the same non-uniformities as batch ALD with higher deposition rate.
C. Optical Properties
To investigate the optical properties of the thin films, three samples where coated and spectrally characterized as depicted in Figs. 6(a)–6(c). A thin single layer of ${{\rm Ta}_2}{{\rm O}_5}$ on quartz substrate with 100 rpm is shown in Fig. 6(a). The transmission measurement (black) depicted in this figure was modeled (green curve), including absorption, and reached high accordance with the measurement. From this curve, a refractive index of 1.98 at 1064 nm and a thickness of 186.2 nm were calculated. This leads to a growth rate of 0.13 nm/sec and a growth per cycle (GPC) of 0.079 nm. For the test wavelength of 1064 nm, an absorption value of 3.1 ppm corresponding to an extinction coefficient of ${1.66} \times {{10}^{- 6}}$ was measured.
Furthermore, a thin ${{\rm SiO}_2}$ coating with 100 rpm was performed. To attain a reliable measurement for the coating on an optical substrate, quartz substrates were coated with 317 nm ion beam sputtered ${{\rm Ta}_2}{{\rm O}_5}$. The corresponding spectra are depicted in Fig. 6(b). A thickness of 192.9 nm was calculated for a refractive index of 1.44 at 1064 nm. This thickness corresponds to a growth rate of 0.13 nm/sec or a GPC of 0.075 nm. Finally, a thick ${{\rm SiO}_2}$ coating was deposited as well on a third substrate with 200 rpm. For modeling of the refractive index and thickness from transmission measurements, again a quartz substrate was precoated (ion beam sputtered) with ${{\rm Ta}_2}{{\rm O}_5}$ and afterwards coated with rALD. A refractive index of 1.44 and a thickness of 1032 nm were calculated. This corresponds to a growth rate of 0.18 nm/sec or a GPC of 0.55 nm. The spectrum of the ${{\rm SiO}_2}$ coating directly deposited on a quartz substrate is depicted also in Fig. 6(c) (magenta). The laser calorimetric absorption measurement leads to a value of 6.0 ppm corresponding to an extinction coefficient of ${4.8} \times {{10}^{- 7}}$ for this sample [14].
4. CONCLUSIONS
It is proven that the presented high-speed ALD coating process is suitable for the production of optical coatings with high quality. High growth rates of up to 0.18 nm/sec are comparable to standard IBS processes and thus enable an efficient fabrication of multilayer systems and thick layer coatings in reasonable time. Furthermore, this technique enables unique properties restricted to the ALD process. Optical online monitoring was successfully realized. It can be expected that this will enable precise coatings and fast process development. Electron microscopic investigation has shown highly defect-free surface and low particle number, both elemental properties for a thin-film coating technique used for high-performance laser coatings. EDX elemental analyses have shown no relevant material concentrations besides coating and substrate materials. Excellent extinction coefficients of ${1.66} \times {{10}^{- 6}}$ for tantalum pentoxide and ${4.8} \times {{10}^{- 7}}$ for silicon dioxide for 200 nm thick single layers at 1064 nm test wavelength were achieved. Experiments on 200 mm diameter silicon wafers have proven low non-uniformity for both coating materials for reasonable areas of up to 140 mm diameter. Higher uniformity is reached in smaller rectangular areas. This enables coating of large substrates as well as a high number of smaller standard substrates. Over the whole 200 mm diameter wafer, comparable non-uniformities as in batch ALD were reached. Lower non-uniformities can be reached with lower rotation speed, but result in decreasing deposition rate [14]; this contradicts the need of high deposition rates for ALD technique. More promising are improvements by modifying the precursor distribution and amount of precursor supply, which will be part of continuous improvement of this still young technique. In near future, coating of microstructures, further coating materials, and more complex coatings like antireflection or high-reflecting coatings and optical filters will become the focus of this research.
Funding
European Regional Development Fund (ZW 2-80147672); and the German Federal Ministry of Education and Research (INTEGRA 01QE2032B); Deutsche Forschungsgemeinschaft (448756425, PhoenixD (EXC 2122, Project ID 390833453).
Acknowledgment
This work was supported by the European Regional Development Fund and also by the German Federal Ministry of Education and Research in the research project. The authors would like to thank the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) for funding this work under Germany’s Excellence Strategy within the Cluster of Excellence PhoenixD.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
REFERENCES
1. P. O. Oviroh, R. Akbarzadeh, D. Pan, R. A. M. Coetzee, and T.-C. Jen, “New development of atomic layer deposition: processes, methods and applications,” Sci. Technol. Adv. Mater. 20, 465–496 (2019). [CrossRef]
2. Y. Li, Y. Xiong, H. Yang, K. Cao, and R. Chen, “Thin film encapsulation for the organic light-emitting diodes display via atomic layer deposition,” J. Mater. Res. 35, 681–700 (2020). [CrossRef]
3. J. Rönn, W. Zhang, A. Autere, X. Leroux, L. Pakarinen, C. A. Ramos, A. Säynätjoki, H. Lipsanen, L. Vivien, E. Cassan, and Z. Sun, “Ultra-high on-chip optical gain in erbium-based hybrid slot waveguides,” Nat. Commun. 10, 1 (2019). [CrossRef]
4. J. Rönn, J. Zhang, W. Zhang, Z. Tu, A. Matikainen, X. Le Roux, E. Durán-Valdeiglesias, N. Vulliet, F. Boeuf, C. A. Ramos, H. Lipsanen, L. Vivien, Z. Sun, and E. Cassan, “Erbium-doped hybrid waveguide amplifiers with net optical gain on fully industrial 300 mm silicon nitride photonic platform,” Opt. Express 28, 27919–27926 (2020). [CrossRef]
5. H. Savin, P. Repo, G. von Gastrow, P. Ortega, E. Calle, M. Garín, and R. Alcubilla, “Black silicon solar cells with interdigitated back-contacts achieve 22.1% efficiency,” Nat. Nanotechnol. 10, 624–628 (2015). [CrossRef]
6. G. S. George, “Atomic layer deposition: an overview,” Chem. Rev. 110, 111–131 (2010). [CrossRef]
7. R. L. Puurunen, “Surface chemistry of atomic layer deposition: a case study for the trimethylaluminum/water process surface chemistry of atomic layer deposition: a case study for the trimethylaluminum/water process,” J. Appl. Phys. 97, 121301 (2005). [CrossRef]
8. K. Pfeiffer, U. Schulz, A. Tünnermann, and A. Szeghalmi, “Antireflection coatings for strongly curved glass lenses by atomic layer deposition,” Coatings 7, 118 (2017). [CrossRef]
9. P. Poodt, D. C. Cameron, E. Dickey, S. M. George, V. Kuznetsov, G. N. Parsons, F. Roozeboom, G. Sundaram, and A. Vermeer, “Spatial atomic layer deposition: a route towards further industrialization of atomic layer deposition,” J. Vac. Sci. Technol. A 30, 010802 (2012). [CrossRef]
10. L. O. Jensen, H. Mädebach, J. Maula, K. Gürtler, and D. Ristau, “Laser induced damage threshold and optical properties of TiO2 and Al2O. coatings prepared by atomic layer deposition,” in Laser-Induced Damage in Optical Materials (2012).
11. P. Poodt, A. Lankhorst, F. Roozeboom, K. Spee, D. Maas, and A. Vermeer, “High-speed spatial atomic-layer deposition of aluminum oxide layers for solar cell passivation,” Adv. Mater. 22, 3564–3567 (2010). [CrossRef]
12. H. Ehlers, S. Schlichting, C. Schmitz, and D. Ristau, “Adaptive manufacturing of high-precision optics based on virtual deposition and hybrid process control techniques,” Chin. Opt. Lett 8, 62–66 (2010). [CrossRef]
13. ISO Central Secretary: Optics and Photonics–Lasers and Laser-Related Equipment–Test Method for Absorptance of Optical Laser Components Version: 2019 (International Organization for Standardization, 2019).
14. L. Kochanneck, A. Tewes, G. A. Hoffmann, K. Niiranen, J. Rönn, H. Velasco, S. Sneck, A. Wienke, and D. Ristau, “Study of rotary atomic layer deposition for optical applications,” Proc. SPIE 12002, 120020D (2022). [CrossRef]
