Fabrication and evaluation of a subwavelength grating in diamond, designed to reduce the Fresnel reflection, is demonstrated. The antireflection (AR) structures are designed to reduce the surface reflection at an illuminating wavelength of 10.6 µm. With this AR-treatment, where no other material is introduced (i.e., no thin film coating), the unique properties of diamond can be fully used. The fabricated AR structures were optically evaluated with a spectrophotometer. The transmission through a diamond substrate with AR structures on both sides was increased from 71% to 97%, with a theoretical value of 99%. Microlenses in diamond are also demonstrated. The lenses are evaluated with interferometers and show good performance. The micro-optical structures were fabricated by electron-beam lithography or photolithographic methods followed by plasma etching.
©2003 Optical Society of America
As an optical material diamond has some unique attributes. It has the widest spectral band of any known material, extending from the ultraviolet into the far infrared region. Its extreme hardness, high thermal conductivity and chemical inertness make diamond an ideal window material for high-power laser applications. However, because of its hardness it becomes very difficult to machine diamond. There have been different attempts of microstructuring diamond, for instance laser ablation [1–3], ion beam milling [4,5] and by a silicon replication method . All of these processes are rather slow and not suitable for more than laboratory demonstrations. However, plasma etching of diamond in high-density systems, which yield high etch rates of diamond, have been presented in the literature [7,8]. Plasma etching is a common technique in the micro-electronic industry for low-cost batch-wise production of semiconductor devices. Our goal has been to find a rapid process which can comply with industrial requirements and still give the high fidelity needed for optical, electronic or biochemical components. Our process for microstructuring diamond is based on three steps: coating diamond substrates with a polymer layer, defining patterns in the layer (by photolithography or some embossing method, for example), and finally transfering the microstructures into diamond by plasma etching. It is vital that the patterns which have been defined in the polymer layer retain their structural fidelity while transfering them into diamond. A thorough investigation of the plasma transfer process has therefore been undertaken.
Two different types of micro-optical elements were designed, fabricated and optically evaluated: an array of spherical microlenses and subwavelength gratings for antireflection (AR) at an illuminating wavelength of 10.6 µm. The reason we chose this particular wavelength is because of the potential of using diamond optics in high-power laser applications together with CO2 lasers. We have earlier, with the technique described in this paper, demonstrated diffractive optical elements in diamond for use together with high power lasers showing good performance . Today, zinc selenide is the material of choice but both theoretical calculations and experimental results show that diamond is superior compared to zinc selenide [10,11].
2. Design and fabrication
The spherical microlenses were fabricated in the following way: a 7 µm thick layer of photoresist was spin-coated on a diamond substrate; using photolithography we next realized cylindrical pillars in the resist; placing the substrate on a hot-plate at 150° C for 10 min. made the cylinders melt which, due to surface tension, resulted in almost perfect spherical caps . Since the base diameter (90 µm) remains the same in the melting step and the volume is constant the height of a lens was about 13 µm.
The subwavelength grating was designed for reducing surface reflections using a commercial optics computer program (GSOLVER, Grating Solver Development Company, USA). This program uses algorithms that solve the vector Maxwell equations in the grating region. The algorithms are based on rigorous coupled wave methods to solve for the interlayer boundary conditions. The program minimizes the surface reflections by using the grating depth, fill factor and grating period as free parameters. The short period of the grating will result in a modulation of the effective refractive index as light passes through the grating, so by tailoring the grating period, grating depth and fill factor the Fresnel reflections due to a difference in the refractive index of diamond (n=2.38 at 10.6 µm) and air can be suppressed, see Ref.  for a general discussion on subwavelength gratings.
The maximum period Λmax, when the incident light is perpendicular to the grating surface, that allows only propagation of the zeroth diffraction order is given by
where λ is the free space wavelength and n is the refractive index of the grating material, assuming that the surrounding medium is air. The maximum period for a diamond AR grating, designed for a wavelength of 10.6µm, is ~4.4 µm. A 2-D subwavelength grating formed by binary square shaped structures were used to reduce the surface reflection. A period of 4 µm was chosen and by varying the duty cycle, which in this case is the width of the binary structure divided with the grating period (which also is identical to the filling factor for binary AR structures), and grating depth an optimal solution can be calculated. This was done with help of GSOLVER, and with a duty cycle of 0.6 and a grating depth of 1.81 µm the calculated Fresnel reflection from one diamond surface was reduced from 17% to 0.3% at a wavelength of 10.6 µm. In Fig. 1 a plot of the calculated transmission through one AR structured diamond surface versus the grating depth, with a duty cycle of 0.6 and an illuminating wavelength of 10.6 µm, is shown. At depth errors smaller than about ±12% the transmission is still above 99% through one surface, but at larger depth errors the transmission drops quickly. Simulating the performance of the AR-grating at various angles of incidence showed that at an angle of 22° the transmission through one surface was 99% and at an angle of 50° the transmission dropped to 95%.
The subwavelength grating was prepared as follows: a 125 nm thin aluminum film was first sputtered on top of a diamond substrate. We also tried depositing aluminum by evaporation but this resulted in poor adhesion between the metal and diamond, which makes the further processing quite uncertain. This fact also makes lift-off processing difficult which otherwise could be a feasible alternative to the method we used. Next, we spin-coated a 180 nm thin polymethylmethacrylate (PMMA) layer on top of the aluminum film, and structured a two-dimensional binary grating in the PMMA layer by electron-beam lithography. We then used PMMA as an etch mask in an inductively coupled plasma (ICP) etching system to open up the aluminum. Etch parameters were ICP power 500 W, bias 10 V, 5 mTorr chamber pressure, flow rates of 45 sccm BCl3 and 5 sccm Cl2, with a total etch time of 5 min. This process was very stable and allowed a well-controlled pattern transfer from PMMA to Al.
3. Plasma etching of the diamond micro-optical structures
Both the spherical lenses and the subwavelength grating were then transfered into the underlying diamond substrates. In both cases we used an oxygen plasma in another ICP etching system. The advantage of using ICP over other etch systems, such as reactive ion etching (RIE), is that ICP gives a high ion density and therefore short etch times. ICP systems also yield better anisotropy, due to low process pressure, and smoother etched surfaces than RIE. Carbon will easily form volatile compounds with oxygen radicals and oxygen plasma is therefore suitable for diamond etching. Etch parameters were ICP power 600W, bias -140V, chamber pressure 2.5 mTorr, flow rates of 7 sccm O2 and 8 sccm Ar, with total etch times of 10– 14 minutes. All samples were mounted with vacuum grease on the water cooled aluminum rf-chuck to enhance the thermal conductivity (e.g., avoid burning of the resist).
The diamond substrates we used were commercial polycrystalline diamond grown by chemical vapour deposition in microwave plasma (Drukker International B.V., The Netherlands). Size of the diamond substrate was 0.3×ϕ 5 mm, flatness < 1 fringe at λ=633 nm. The surfaces were polished to a root-mean-square (RMS) roughness below 15 nm.
By first measuring the etch rate of partly covered diamond and knowing the desired grating depth, we could calculate the etch time needed for fabricating the subwavelength grating. A silicon wafer with resist on was also partly covered to find the etch rate in the resist and thereby we could calculate the etch time for the transfer of the resist microlenses into diamond. The etch rate of diamond and resist was measured to be 200 nm/min and 1300 nm/min, respectively. These etch rates corresponds to a total etch time of about 10 min. Finally, for the AR structures, the Al was stripped by wet etching. In Fig. 2 a white-light interferometer picture of a diamond microlens is shown and Fig. 3 shows the result of a twodimensional subwavelength grating in diamond. The diamond microlenses and the subwavelength grating were then optically characterized.
4. Optical characterization
The fabricated diamond microlenses were evaluated with a white light interferometer and a Twyman-Green interferometer. The focal length of the fabricated diamond lenses is 365 µm, and the f-number is 4. The deviation from a perfect spherical surface was measured to be 6.7 nm (rms) over more than 90% of the lens area. These values were calculated by using data, generated by the white light interferometer, from a 2-D surface scan of the lens surface. The phase error, measured with a Twyman-Green interferometer at λ=633 nm, was found to be less then 31 nm (λ/20), see Fig. 4. Both of these results verify the high fidelity of the manufacturing process. In our case, the selectivity did not equal one and therefore a spherical resist surface will not imply a spherical diamond surface, but rather an ellipsoidal curve. However, since the diamond lens is so shallow the difference between a spherical and ellipsoidal surface is small.
The AR structured diamond substrates was evaluated in the IR-region. The grating period is 4 µm and the size of each square is 2.4×2.4µm2. The grating depth is 1.8 µm. This two-dimensional design of AR-structures is valid for both unpolarized and polarized light at normal incidence with a calculated transmission of 99.5% for double-sided treatment. Since an optical component has two interfaces relative its surrounding medium, we fabricated such subwavelength gratings on both sides of a diamond substrate. Figure 5 shows the optical transmission for unstructured diamond, as well as one- and two-sided AR-coated diamond measured with a spectrophotometer.
In the transmission measurements we used a pinhole with a diameter of 2.4 mm to reduce the beam width, since the size of the AR-structured area is 2.5×2.5 mm2. The calculated transmission values of the one- and two-sided AR-coated diamond are also shown in Fig. 5. As can be seen the transmission is increased from 70.8% (maximum theoretic transmission is 71.4%) for the blank diamond to 81,3% (theoretic maximum value of a perfect AR-coating is 83.3%, calculated maximum value using GSOLVER is 82.9%) of one-side treated diamond and finally 96.6% (theoretic maximum value is 100%, calculated maximum value is 99.5%) for a double-side treated diamond substrate. The main reason that the calculated value is not reached is probably due to stitching errors in the e-beam lithography system used. The ARsurface is written in an array of 10×10 identical sub-patterns and due to the poor stage control we induce a lateral error of ~2 µm every time the stage is moved. Other possible error sources are wrong etch depth or lack of control of the feature size, the estimated control of etch depth and feature size is in the range of +/- 200 nm.
5. Discussion and conclusions
Clearly, we have been able to drastically increase the optical transmission of diamond, which due to its relatively high refractive index otherwise might prove to be a problem in optical applications. By avoiding thin film coating of various materials to reduce the surface reflection of diamond, full advantage of diamond’s unique properties can be taken. Both binary and continuous-relief micro-optical structures in diamond have been demonstrated in this paper. The optical structures have been optically evaluated and showed excellent performance. In this study we used e-beam lithography to define the AR pattern but standard photolithography with high enough accuracy can be used to meet industry requirements. The etch rate of diamond was found to range from 150 to 250 nm/min with possibilities to obtain a selectivity between diamond and photoresist of 0.15–0.2. This means that with a structure depth of 100 µm in photoresist the maximum etched depth in diamond would be 20 µm. To obtain even deeper structures one could use another type of polymer material with a lower etch rate than photoresist, and emboss the microstructures into such a polymer material rather than first defining them by lithographic techniques. Such an approach is also more appealing from a manufacturing standpoint since embossing is a faster process than lithography. The ICP etching produced very smooth surfaces with low surface roughness, less than 15 nm (rms). This is in the same range as for the untreated substrates. An evaluation of fabricating micro-optical structures on different sides of the diamond wafers showed that micromachining the side with a low density of grain boundaries having larger grain size (the side which has been closest to the seed layer during the CVD fabrication) gives a slightly better result in terms of surface roughness. The ideal case would be to use monocrystalline diamond, but since this is still typically much too expensive for most applications we found it worthwhile to investigate polycrystalline diamond for our micromachining method. One can expect the results to be even better using monocrystalline diamond since this will give less light scattering and lower surface roughness. With the ability to fabricate different shapes of microstructures in diamond other applications such as electronic [14,15] and biomedical [16–17] devices can also start to be explored.
The authors gratefully acknowledge Jacob Jonsson at Uppsala University for help with the spectrophotometer equipment and Herman Högström at Uppsala University for help with the fabrication of the microlenses. This work was in part financed by SUMMIT, the Swedish Center for Surface and Microstructure Technology, supported by the Swedish Agency for Innovation Systems (VINNOVA).
References and links
1. S. S. M. Chan, F. Raybould, G. Arthur, F. Goodall, and R. B. Jackman, “Laser projection patterning for the formation of thin film diamond microstructures,” Diamond and Related Mater. 5, 317–320 (1996). [CrossRef]
2. J. D. Hunn, S. P. Withrow, C. W. White, R. E. Clausing, L. Heatherly, and C. P. Christensen, “Fabrication of single-crystal diamond microcomponents,” Appl. Phys. Lett. 65, 3072–3074 (1994). [CrossRef]
3. S. Gloor, V. Romano, W. Luthy, H. P. Weber, V. V. Kononenko, S. M. Pimenov, V. I. Konov, and A. V. Khomich, “Antireflection structures written by excimer laser on CVD diamond,” Appl. Phys. A 70, 547–550 (2000). [CrossRef]
4. M. Tarutani, Y. Takai, and R. Shimizu, “Application of the focused-ion-beam technique for preparing the cross-sectional sample of chemical vapor deposition diamond thin film for high-resolution transmission electron microscope observation,” Jap. J. Appl. Phys. 31, 1305–1308 (1992). [CrossRef]
5. E. G. Spencer and P. H. Schmidt, “Ion machining of diamond,” J. Appl. Phys. 43, 2956–2958 (1972). [CrossRef]
6. V. G. Ralchenko, A. V. Khomich, A. V. Baranov, I.I. Vlasov, and V. I. Konov, “Fabrication of CVD diamond optics with antireflective surface structures,” Phys. Status Solidi A 174, 171–176 (1999). [CrossRef]
7. S. J. Pearton, A. Katz, F. Ren, and J. R. Lothian, “ECR plasma etching of chemically vapour deposited diamond thin films,” Electron. Lett. 28, 822–824 (1992). [CrossRef]
8. M. Karlsson, K. Hjort, and F. Nikolajeff, “Transfer of continuous-relief diffractive structures into diamond by use of inductively coupled plasma dry etching,” Opt. Lett. 26, 1752–1754 (2001). [CrossRef]
9. M. Karlsson and F. Nikolajeff, “Fabrication and evaluation of a diamond diffractive fan-out element for high power lasers,” Opt. Express 11, 191–198 (2003). http://www.opticsexpress.org/abstract.cfm?URI=OPEX-11-3-191 [CrossRef] [PubMed]
10. B. Dischler and C. Wild, Low-Pressure Synthetic Diamond (Springer, Berlin Heidelberg, 1998). [CrossRef]
11. C. J. Brierley, C. M. Beck, G. R. Kennedy, J. Metcalfe, and D. Wheatley, “The potential of CVD diamond as a replacement for ZnSe in CO2 laser optics,” Diamond and Related Mater. 8, 1759–1764 (1999). [CrossRef]
14. J. Isberg, J. Hammersberg, E. Johansson, T. Wikstrom, D. J. Twitchen, A. J. Whitehead, S. E. Coe, and G. A. Scarsbrook, “High Carrier Mobility in Single-Crystal Plasma-Deposited Diamond,” Science 297, 1670–1672 (2002). [CrossRef] [PubMed]
16. C. E. Troupe, I. C. Drummond, C. Graham, J. Grice, P. John, J. I. B. Wilson, M. G. Jubber, and N. A. Morrison, “Diamond-based glucose sensors [diabetic blood analysis],” Diamond and Related Mater. 7, 575–580 (1998). [CrossRef]
17. M. Adamschik, M. Hinz, C. Maier, P. Schmid, H. Seliger, E. P. Hofer, and E. Kohn, “Diamond micro system for bio-chemistry,” Diamond and Related Mater. 10, 722–730 (2001). [CrossRef]