Over the last 15 years, breakthroughs in the manufacture and processing of diamond grown by chemical vapor deposition (CVD) have established diamond as an excellent substrate material for high-power and high-energy optics. Diamond is a natural choice for these highly demanding applications due to a combination of desirable properties including: extremely broad transmission spectrum, low absorption, chemical inertness, mechanical strength, and the highest room temperature thermal conductivity of any material. These properties allow diamond to perform in environments and applications where other materials are simply not viable options.
Broadly, there are two types of CVD diamond — single crystal (SC) and polycrystalline (PC) — each with their distinct advantages. Single crystal diamond possesses the lowest absorption, scatter, and birefringence, whereas polycrystalline diamond can be manufactured and processed to sizes of up to 135 mm diameter.
For many laser applications there is an industry trend of continually increasing power and energy, both total value and density. One example is EUV pumping systems which use CO2 lasers at 10.6 μm wavelength. These pumping systems currently utilize continuous wave (CW) power levels of 20kW, a number which is expected to double over the next few years (EUV Litho Symposium, 2013). Diamond’s combination of thermal and optical properties make it a unique material capable of handling these extreme power levels.
That said, the same is not always true for thin film anti-reflection coatings needed to limit potentially damaging back reflections and maximize throughput energy. In many instances, the thermal properties of diamond are 1000x better than those of the AR coating materials applied to the diamond surface. As a result, laser induced damage threshold (LIDT) of coated diamond optics is much more dependent on the AR coating materials than on the diamond substrate. This is reflected in the observation that failures of diamond windows in the field, which can be very costly, are almost always related to failure of the thin film coating. As industry pushes towards higher power and energy, the reliability of thin film coatings presents a major challenge.
To meet this challenge, Element Six, in collaboration with Harvard University’s John A. Paulson School of Engineering and Applied Sciences, has developed an anti-reflective solution that eliminates the thin film coating altogether. Available now for the CO2 laser wavelength 10.6 μm, the new Diamond PureOptics™ product uses 3D microstructures etched into the window surfaces to achieve anti-reflective properties without introducing inferior non-diamond materials. The Diamond PureOptics™ manufacturing process, which uses standard semiconductor equipment, is easily scalable and similar in cost to thin film coatings. The AR performance of these so called meta-surfaces as shown (Figure 1) is at least as good, or better, than traditional thin films.
This all-diamond solution has many benefits over traditional thin film AR coatings. First, as shown in Figure 2, reliability is increased by a factor of at least 10x over traditional AR coating solutions in regards to laser induced damage. In tests of CW LIDT @ 10.6 μm the Diamond PureOptics™ have gone undamaged, even at power densities up to 3 MW/cm2. In addition to increased LIDT, optical absorption of the window is decreased due to the absence of thin film coating layers which typically absorb 0.1% per surface. This has the valuable benefit of reducing heat driven effects such as thermal lensing. Also, as the Diamond PureOptics™ windows do not possess a coating-substrate interface in the traditional sense, there is no CTE (coefficient of thermal expansion) mismatch between layers, no shear stress, and it is impossible for the meta-surfaces to delaminate. Lastly, due to the chemically inert nature of diamond, the AR meta-surfaces boast unparalleled environmental stability and can be easily and thoroughly cleaned using aggressive acids and solvents.
Effective Medium Theory
Anti-reflective coatings on diamond are typically made by depositing a dielectric stack of materials with refractive indices and thicknesses to ensure proper destructive interference. For an ideal single layer AR coating, refractive index and thickness are found using the formula seen in Figure 4.
Such heterostructures display a decreased LIDT compared to bulk diamond. However, bulk diamond can be engineered to display an effective index equal to the required anti-reflective index condition. This can be done by etching a periodic structure in the diamond with a periodicity satisfying the formula seen in Figure 5.
Where λ0 is the wavelength of interest. For a wavelength of 10.6 μm, this corresponds to a periodicity of 4.4 μm. Seen from the point of view of diffractive optics, so long as any structures fabricated on the diamond satisfy this requirement, incident light will only diffract into the zeroth order[1].
In real structures, such as the truncated cones illustrated in Fig. 3(a), the sidewalls are not necessarily straight. This is advantageous in two ways, by reducing the index contrast at the cone apex and bottom and by providing a spatially-varying refractive index which minimizes back-reflections[2][3]. The refractive index profile along such a cone is shown schematically in Fig. 3(b). To estimate the refractive index at a particular z-value, we can employ the Bruggeman approximation[4][5] seen in Figure 6.
Here, f denotes the filling fraction of the etched pattern (i.e. ratio of structure area to period area), and g denotes a geometry-dependent depolarization factor which we approximate as 0.5 for infinitely long ellipsoids[5]. Using this formula, we see that the ideal filling fraction is f=0.5, meaning that half of etched surface is occupied by the diamond structures.
Because of the spatially varying refractive index, this approximation is not strictly correct. Although a solution using an effective medium approach can be derived by using a transfer matrix method or rigorous-coupled-wave analysis (RCWA)[22], full finite-difference time-domain (FDTD) simulations of the etched substrate can establish the transmission of the structure with complete rigor. Using LUMERICAL, the design parameters were varied to derive the optimal diamond shape. A representative parameter sweep is shown in Figure 3(c), where the pillar height is swept while holding the pillar top and bottom radius constant. This simulation shows an optimal design height of ~ 1.75 μm.
Diamond Processing
Although it is the hardest material known to science, diamond is readily shaped with traditional semiconductor processing techniques. Plasma etching, where a mixture of gases is excited into a plasma and then directed onto the targeted material, is by far the most common method. By using a layer (known as the resist) to block the etching of diamond in certain areas, structures with straight sidewalls can be made. This technique has seen use across a wide variety of applications in diamond, including optical waveguides[6], nanowires for quantum optics[7], and microlenses[1][8].
There are two main plasma chemistries which are used to etch diamond: an argon/chlorine mixture which etches slowly (~80 nm/min) and is used to smooth diamond substrates [9], and a pure oxygen plasma. The etching mechanism of the oxygen plasma is mostly chemical, as excited oxygen species hit the target diamond surface, reacting with surface carbon to form carbon dioxide and other volatile species, quickly leaving the etched surface. This chemistry has the benefit of being faster (~160 nm/min) and more selective to the diamond, preserving the resist material and shape[7].
Additional Applications
Diamond's excellent material properties position it to be a prime player in next-generation optical and mechanical systems. Processing techniques such as Faraday cage-angled etching[10][11][12], masked oxidation[13], quasi-isotropic plasma etching[14], and two-photon assisted etching[15] may enable new paradigms for diamond-based devices. Outside of applications in free-space mid-infrared optics like antireflective coatings[2] and wave plates[16], enormous potential remains to be tapped in the shorter side of the electromagnetic spectrum, provided that fabrication challenges stemming from the required smaller structures can be met. Beyond free-space optics, diamond may play a larger role in niche integrated optics devices, where optical components are fabricated directly on a silicon chip. Examples include integrated frequency combs[17], filters[18], Raman lasers[19], and quantum optical networks [20].
This article was written by Pawel Latawiec, Alexander Muhr, Marko Loncˇar, and Daniel Twitchen of Element Six (Santa Clara, CA). For more information, contact Daniel Twitchen at
References
- M. Karlsson and F. Nikolajeff, Optics express 11, 502 (2003).
- A. B. Harker and J. F. Denatale, Proc. SPIE 1760, 261 (1992).
- D. S. Hobbs and B. D. MacLeod, Proc. SPIE 5786, 349 (2005).
- D. A. G. Bruggeman, Ann. Phys. 24, 636 (1935).
- H. Wang, X. Liu, L. Wang, and Z. Zhang, Proc. SPIE 8495, 62 (2013).
- M. P. Hiscocks, K. Ganesan, B. C. Gibson, S. T. Huntington, F. Ladouceur, and S. Prawer, Optics Express 16, 19512 (2008).
- B. J. M. Hausmann, M. Khan, Y. Zhang, T. M. Babinec, K. Martinick, M. McCutcheon, P. R. Hemmer, and M. Loncar, Diam. Relat. Mater. 19, 621 (2010), arXiv:0908.0352.
- H. W. Choi, E. Gu, C. Liu, C. Griffin, J. M. Girkin, I. M.Watson,and M. D. Dawson, J. Vac. Sci. Technol. B 23, 130 (2005).
- H. A. Atikian, A. Eftekharian, A. Jafari Salim, M. J. Burek, J. T. Choy, A. Hamed Majedi, and M. Loncar, Appl. Phys. Lett. 104, 22 (2014), arXiv:1401.4490.
- M. J. Burek, N. P. De Leon, B. J. Shields, B. J. M. Hausmann,Y. Chu, Q. Quan, A. S. Zibrov, H. Park, M. D. Lukin, and M. Loncar, Nano Lett. 12, 6084 (2012).
- K. K. Kovi, R. S. Balmer, and J. Isberg, Diam. Relat. Mater. 58, 185 (2015).
- P. Latawiec, M. J. Burek, Y.-I. Sohn, and M. Loncar, J. Vac. Sci. Technol. B 34, 041801 (2016), arXiv:1603.03735.
- C. D. McGray, R. A. Allen, M. Cangemi, and J. Geist, Diam. Relat. Mater. 20, 1363 (2011).
- B. Khanaliloo, M. Mitchell, A. C. Hryciw, and P. E. Barclay, Nano Lett. 15, 5131 (2015).
- A. Lehmann, C. Bradac, and R. P. Mildren, Nature Comms. 5, 3341 (2014).
- C. Delacroix, P. Forsberg, M. Karlsson, D. Mawet, O. Absil, C. Hanot, J. Surdej, and S. Habraken, Applied optics 51, 5897 (2012).
- B. J. M. Hausmann, I. Bulu, V. Venkataraman, P. Deotare, and M. Loncar, Nature Photonics 8, 369 (2014).
- B. J. M. Hausmann, I. B. Bulu, P. B. Deotare, M. McCutcheon, V. Venkataraman, M. L. Markham, D. J. Twitchen, and M. Loncar, Nano Letters 13, 1898 (2013).
- P. Latawiec, V. Venkataraman, M. J. Burek, B. J. M. Hausmann, I. Bulu, and M. Loncar, Optica 2, 924 (2015).
- B. J. M. Hausmann, B. Shields, Q. Quan, P. Maletinsky, M. McCutcheon, J. T. Choy, T. M. Babinec, A. Kubanek, A. Yacoby, M. D. Lukin, and M. Loncar, Nano Lett. 12, 1578 (2012), arXiv:1111.5330v1.
- H. Kikuta, Y. Ohira, and K. Iwata, Applied Optics 36, 1566 (1997).
- I. Richter, P. C. Sun, F. Xu, and Y. Fainman, Applied Optics34, 2421 (1995).