We report on factors affecting the performance of a broadband, mid-IR absorber based on multiple, alternating dielectric / metal layers. In particular, we investigate the effect of interface roughness. Atomic layer deposition produces both a dramatic suppression of the interface roughness and a significant increase in the optical absorption as compared to devices fabricated using a conventional thermal evaporation source. Absorption characteristics greater than 80% across a 300 K black body spectrum are achieved. We demonstrate a further increase in this absorption via the inclusion of a patterned, porous anti-reflection layer.
© 2014 Optical Society of America
There is presently great interest in creating structures which exhibit high levels of optical absorption over broad spectral ranges for applications including energy scavenging [1–4]. A practical, high efficiency broadband absorber would find applications in the collection of radiant energy, acting as the front end of energy harvesting devices from remote or nearby heat sources. A specific example of such an application of such a structure could be for harvesting of infrared (IR) radiation from the sun or the earth (“earthshine”), to provide electrical power to a satellite in space. A second application would be as an artificial source of broadband radiation: indeed the basis for selective thermal radiation lies in the ability to control a material’s absorptivity which by Kirchoff’s law is equivalent to controlling the material emissivity . We note that there has been a long history of the design and/or fabrication of structures for tailoring the reflectance and absorption of light in various parts of the spectrum. Such structures include periodic metal gratings; these however tend to be narrow band and produce both directional and polarization dependent absorption [6,7]. There have also been a number of recent studies, both theoretical and experimental on the use of “metamaterial” structures as perfect or near-perfect absorbers; these are typically bandwidth limited or require multiple layers whose relative positions must be precisely controlled to optimize performance [8–17]. Carbon nanotube “forests” have recently been shown to outperform more traditional, highly porous “blacks,”  however coatings of each of these types of are quite fragile. Quarter-wavelength structures for thermal detectors, based upon the need to minimize the thermal mass, and thus keep the number of layers to a minimum produce less than optimal performance [19–22]. Very recent reports show that similar structures can be achieved using layers whose thickness can be reduced well below a quarter wavelength using strong interference effects using highly absorbing media  or structures consisting of a thin high index layer on a heavily doped semiconductor ; the structures reported on in these cases however demonstrated only relatively narrow bands of near-perfect absorption.
In recent work we carried out calculations of the performance of an idealized, near-perfect blackbody absorber based upon a new approach: a multilayer structure consisting of dielectric layers interspersed with thin metallic layers [4, 24]; such structures are appealing for their ease of fabrication and for potential near perfect broadband absorption. Metallic layers had previously been used for tailored-absorption devices [19–21], but the absorption in thick metal layers is accompanied by significant reflection and/or transmission, reducing the absorption efficiency. Our calculations  were based upon multiple thin NiCr metallic layers, each of which absorbs a small fraction of the radiation, alternating with thick BaF2 dielectric layers (which are transparent over the mid-IR) predicted that absorption close to 100% might be achieved over a broad band at wavelengths between Bragg resonances, where the center wavelength can be tuned by changing the dielectric layer thickness. However, these calculations were made for a defect-free multilayer structure, specifically one without interface roughness. In real fabricated devices, this roughness cannot be ignored. Although roughness can be included in the multilayer simulation, a unique solution is difficult to obtain because a variety of different individual layer properties can result in the same solution.
Here we investigate experimentally the mid-IR absorbance of fabricated multilayer structures designed to match the emission spectrum of a Blackbody near room temperature. We find that the interface roughness resulting from conventional thin film deposition strongly degrades the absorption spectrum and that reducing the roughness using atomic layer deposition (ALD) leads to performance approaching that calculated for an ideal structure. Finally, we demonstrate that a significantly improved performance is afforded by adding a patterned anti-reflection layer on such a structure.
We begin with a brief description of the experiments. A schematic of the multilayer structure used in our initial investigations is shown in Fig. 1(a). In fabricating this, first we deposited 100 nm of Ag from an evaporative source onto Si(001) substrates at room temperature. This “bottom layer” blocks transmission of any IR radiation that might make it through the upper absorptive layers. Alternating films of BaF2, 1.8 um thick, and NiCr (80% Ni, 20% Cr by weight), 2 nm thick were then deposited onto this bottom layer using electron-beam irradiated sources in high vacuum. In these multilayer structures absorption minima occur at wavelengths corresponding to Bragg reflection resonances. For the second type of device discussed below, we instead used a Beneq TFS 500 ALD system to grow alumina, utilizing trimethyl aluminum (TMA) and H2O precursors at a temperature of 150°C. This resulted in a growth rate per cycle of approximately 0.96 Å/sec. We deposited 2 nm Ti layers and a 100 nm Al bottom layer by e-beam evaporation. We note that it should be possible to grow Ti by ALD  to suppress roughness further. The surface morphology of both types of devices was measured using atomic force microscopy (AFM) in tapping mode. We used a Bomem FTIR Spectro-radiometer to measure the reflectivity spectrum from our structures at a fixed incident angle of 5°. Appropriate sources, beam splitters and detectors were chosen to cover a range of wavelengths from 2 μm to 80 μm. For the specific example mentioned in the introduction of a satellite harvesting earthshine we note that 99.6% of the total intensity for blackbody of 300 K is emitted in this range.
3. Results and discussion
In Fig. 2 we compare the calculated absorption spectra based on a multiple reflection Fresnel analysis  with those determined experimentally. We can experimentally determine emissivity by accounting for all of the light not absorbed by the material. A material with sub-wavelength surface roughness and geometry can be approximated as homogeneous, thus removing scattering effects. With the inclusion of an optically thick bottom layer of metal (silver in this case) as the material ground plane preventing transmission, Kirchoff’s law reduces to the simplified form:Figure 2(a) shows a comparison for devices including 2 periods (red dashed curve vs. blue solid curve) and 10 periods (gold dashed curve vs. green solid curve) of alternating BaF2 and NiCr layers. For both devices the measured absorptance spectra show little resemblance to those we calculated. The pronounced Bragg reflection-dips are missing in the experimental spectra, and the experimental absorptivities generally fall significantly below those calculated.
What is the source of the large discrepancy? While there are a number of possible contributing factors, including possible non-stoichiometry of the dielectric and interdiffusion of the thermally grown Nichrome and BaF2 layers, an obvious candidate is interface roughness and the resulting scattering [27,28]. Simple models of conventional thin film growth predict that the interface roughness should increase monotonically as the overall thickness of multilayer coating increases [29,30]. Indeed a comparison for 2 period devices vs. 10 period devices shows a larger discrepancy in the latter case. To further test the hypothesis that roughness induced scattering is responsible for the discrepancy we track the surface roughness of a device as a function of overall thickness using AFM. Results from these measurements for our BaF2/NiCr multilayer devices are summarized in Fig. 3(a) and Fig. 3(c) with the latter panel showing the root-mean-square (RMS) surface roughness vs. the overall device thickness. The RMS roughness amplitude (green dashed curve) is more than 40 nm for a 2-period device, increasing approximately as the square root of the thickness or equivalently the number of periods. The large magnitude of the roughness for these devices makes it unlikely that the 2 nm NiCr layers are contiguous, probably frustrating the Joule-heating dissipation that would be expected in the ideal case.
The roughness which develops during conventional film growth can be almost entirely suppressed by replacing it with ALD . ALD takes advantage of a series of self-limiting reactions to build up a film a monolayer at a time [32,33] with near-exact conformality to the underlying substrate [34,35]. Among the dielectrics which can be grown via ALD, amorphous alumina (Al2O3) is perhaps the easiest to grow with nearly perfect conformality .
To investigate ALD as a method for improving performance, we thus replaced the BaF2 dielectric layers with ALD alumina. We also replaced nichrome with titanium as it is known to wet alumina, shows high resistivity, and avoids variable stoichiometry issues since it is a pure metal. Our initial attempts to grow multiple period structures using these materials resulted in a tendency for the alumina to flake off, seemingly due to the difference in the thermal expansion coefficient between the Si substrate and the Al2O3 layers which are deposited at elevated temperature. We found switching to sapphire substrates alleviated the flaking. In this case we used a (100 nm thick) aluminum (Al) bottom reflection layer, due to its excellent adhesion to sapphire. The overall structure of this device is illustrated schematically in Fig. 1(b).
Indeed, as illustrated by the AFM images of Fig. 3, we find that the very high level of conformality of ALD films results in much smaller interface roughness. The blue solid circles in Fig. 3(c) quantify the roughness as a function of the overall thickness of the multilayer structures. The measured RMS roughness of the sapphire substrates was approximately 8 nm; the surface roughness remains essentially unchanged from this value upon deposition of the ALD alumina/Ti multilayer structures. A result of this smaller roughness is that a finite sheet resistance, on the order of several, is measurable for each of the individual Ti layers. This provides direct confirmation of the presence of contiguous thin metal films between the individual ALD dielectric layers. By contrast the apparent sheet resistance of the nichrome layers on the much rougher thermally evaporated BaF2 was immeasurably high, suggesting that they were not contiguous.
Our results show that the large reduction in interface roughness is accompanied by a dramatic improvement in the optical absorption characteristics of the absorber. Based on the dielectric constant of Al2O3, the thickness of the individual dielectric layers for the ALD-fabricated devices was chosen to be 1.16 μm, giving rise to a Bragg resonance absorption minimum at a wavelength of 3.85 μm. Figure 4(a) and 4(b), shows a comparison between experiment and theory for these structures. For the calculated spectra the complex refractive index of ALD grown Al2O3 was derived empirically using a baseline complex refractive index determined by ellipsometry, which was then modified slightly to optimize the fit in Fig. 4(a) and Fig. 4(b). The resulting complex refractive index is summarized in Fig. 4(c). The agreement between calculated and measured spectra in Fig. 4(a) and Fig. 4(b) is excellent. In particular, the Bragg resonance near 3.85 μm (indicated by the arrow) is clearly evident, and narrows dramatically as the number of Al2O3 periods increases as expected. Most importantly, the experimentally determined absorption integrated over the wavelength range from 4 μm to 40 μm is 76% for a 2 period device and 82% for a 10 period structure (blue dots in Fig. 5(d)). These values far exceed those for the corresponding devices produced by thermal evaporation of BaF2.
To separate the effect of interface roughness from those due to differences in the dielectric, we also fabricated devices using thermally evaporated Al2O3. In a similar manner to devices based on thermally evaporated BaF2, the RMS surface roughness again increases with overall thickness (red triangles in Fig. 3(c)). This led to a smaller overall absorption than that for the ALD-based devices, as can be seen by comparing the spectra in Fig. 2(b) with those in Figs. 4(a), 4(b). In addition, the reduced precision in reproducing the Al2O3 layer thicknesses by thermal evaporation results in broader Bragg resonances than for the ALD-based devices.
There is however a disadvantage in the use of Al2O3 as the dielectric material in the multilayer absorber: it is not completely transparent in mid-IR range. A signature of this is that the extinction coefficient K shows a peak at a wavelength of 12.5 μm (Fig. 4(c)). We find that this manifests itself as a dip in our multilayer absorption spectra, or equivalently, a peak in the reflection spectra. An obvious question is whether this effect could be mitigated through the use of an anti-reflection (AR) layer above the multilayer structures.
An ideal, simple antireflection layer  would have a refractive index ofEq. (3) and consisting of a sub-wavelength pattern of pores within an Al2O3 matrix. The dielectric function of such a layer can be approximated by Maxwell–Garnett effective medium theory [37,38] (EMA):
For ease of fabrication and scalability, we attempted to make such a pattern using photolithography. As a rough estimate for the range of applicability of effective medium theory, we designed the pattern period to be less than one quarter of the center wavelength of the absorption band. In general, thickness of the photoresist should be 2-3 times thicker than target thickness before the liftoff process. Based upon these considerations we chose SU8 3005 resist (MicroChem) which supports high aspect ratio fabrication and is inert to most usual chemicals.
Using a combination of AFM and scanning electron microscopy (SEM) to determine dimensions, we were able to fabricate cylindrical pores with diameters of 1.3 μm, lateral spacings of 2.8 μm, and depths of 2.4 μm (Fig. 5(b)); this corresponds to a volume fraction of 0.4. The absorption spectrum measured for the 2 period-structure terminated with this AR layer illustrated schematically in Fig. 1(c) shows a noticeable reduction in the absorption dip near 12.5 μm as can be seen from the blue curve in Fig. 5(c). The resulting overall absorption, integrated over the range 4 μm to 40 μm, is increased by approximately 10% compared to that for an identical structure without this AR layer as shown in Fig. 5(d). Our calculations, summarized by the solid curves in Fig. 5(d), indicate that a similar improvement should be possible for devices consisting of larger numbers of dielectric/metal periods if terminated by such an AR layer.
In summary, we have designed, fabricated, and characterized structures consisting of alternating dielectric and thin metal layers which indeed show strong absorption across a broad band in the mid-IR range. Based upon Kirchoff’s law we expect that these might be employed as midIR emitters over bands considerably broader than that for a blackbody. Our results show that the optical absorption spectra of these devices are strongly affected by interface roughness at the level typical of conventional thin film growth techniques. We demonstrate that the mid-IR absorption in these structures can be significantly enhanced by replacing thermal evaporation-based fabrication with atomic layer deposition, and attribute this to the much smaller interface roughness which results from growth. We demonstrate that an anti-reflection layer consisting of a sub-wavelength patterned porous dielectric layer mitigates the lack of ideal transparency of the dielectric within the mid-IR range. Finally, we expect that even higher levels of optical absorption should be possible if layers of a dielectric which is transparent in the mid-IR, such as BaF2 are deposited via ALD.
We acknowledge useful discussions with T.D. Corrigan.
References and links
1. E. E. Narimanov and A. V. Kildishev, “Optical black hole: Broadband omnidirectional light absorber,” Appl. Phys. Lett. 95(4), 041106 (2009). [CrossRef]
2. K. Aydin, V. E. Ferry, R. M. Briggs, and H. A. Atwater, “Broadband polarization-independent resonant light absorption using ultrathin plasmonic super absorbers,” Nat. Commun. 2, 517 (2011).
4. T. D. Corrigan, D. H. Park, H. D. Drew, S.-H. Guo, P. W. Kolb, W. N. Herman, and R. J. Phaneuf, “Broadband and mid-infrared absorber based on dielectric-thin metal film multilayers,” Appl. Opt. 51(8), 1109–1114 (2012). [CrossRef] [PubMed]
5. W. Streyer, S. Law, G. Rooney, T. Jacobs, and D. Wasserman, “Strong absorption and selective emission from engineered metals with dielectric coatings,” Opt. Express 21(7), 9113–9122 (2013). [CrossRef] [PubMed]
6. J. Le Perchec, P. Quémerais, A. Barbara, and T. López-Ríos, “Why metallic surfaces with grooves a few nanometers deep and wide may strongly absorb visible light,” Phys. Rev. Lett. 100(6), 066408 (2008). [CrossRef] [PubMed]
7. V. G. Kravets, F. Schedin, and A. N. Grigorenko, “Plasmonic blackbody: almost complete absorption of light in nanostructured metallic coatings,” Phys. Rev. B 78(20), 205405 (2008). [CrossRef]
8. C. Wu, B. Neuner, G. Shvets, J. John, A. Milder, B. Zollars, and S. Savoy, “Large-area wide-angle spectrally selective plasmonic absorber,” Phys. Rev. B 84(7), 075102 (2011). [CrossRef]
10. J. Hao, J. Wang, X. Liu, W. J. Padilla, L. Zhou, and M. Qiu, “High performance optical absorber based on a plasmonic metamaterial,” Appl. Phys. Lett. 96(25), 251104 (2010). [CrossRef]
11. Y. Avitzour, Y. A. Urzhumov, and G. Shvets, “Wide-angle infrared absorber based on a negative-index plasmonic metamaterial,” Phys. Rev. B 79(4), 045131 (2009). [CrossRef]
12. M. Diem, T. Koschny, and C. M. Soukoulis, “Wide-angle perfect absorber/thermal emitter in the terahertz regime,” Phys. Rev. B 79(3), 033101 (2009). [CrossRef]
13. N. I. Landy, C. M. Bingham, T. Tyler, N. Jokerst, D. R. Smith, and W. J. Padilla, “Design, theory and measurement of a polarization-insensitive absorber for terahertz imaging,” Phys. Rev. B 79(12), 125104 (2009). [CrossRef]
15. H. Tao, C. M. Bingham, A. C. Strikwerda, D. Pilon, D. Shrekenhamer, N. I. Landy, K. Fan, X. Zhang, W. J. Padilla, and R. D. Averitt, “Highly flexible wide angle of incidence terahertz metamaterial absorber: design, fabrication, and characterization,” Phys. Rev. B 78(24), 241103 (2008). [CrossRef]
16. H. Tao, N. I. Landy, C. M. Bingham, X. Zhang, R. D. Averitt, and W. J. Padilla, “A metamaterial absorber for the terahertz regime: design, fabrication and characterization,” Opt. Express 16(10), 7181–7188 (2008). [CrossRef] [PubMed]
17. T. V. Teperik, F. J. García de Abajo, A. G. Borisov, M. Abdelsalam, P. N. Bartlett, Y. Sugawara, and J. J. Baumberg, “Omnidirectional absorption in nanostructured metal surfaces,” Nat. Photonics 2(5), 299–301 (2008). [CrossRef]
18. K. Mizuno, J. Ishii, H. Kishida, Y. Hayamizu, S. Yasuda, D. N. Futaba, M. Yumura, and K. Hata, “A black body absorber from vertically aligned single-walled carbon nanotubes,” Proc. Natl. Acad. Sci. U.S.A. 106(15), 6044–6047 (2009). [CrossRef] [PubMed]
20. P. Eriksson, J. Y. Andersson, and G. Stemme, “Interferometric, low thermal mass IR-absorber for thermal infrared detectors,” Phys. Scr. T 54, 165–168 (1994). [CrossRef]
22. A. D. Parsons and D. J. Pedder, “Thin-film infrared absorber structures for advanced thermal detectors,” J. Vac. Sci. Technol. A 6(3), 1686–1689 (1988). [CrossRef]
23. M. A. Kats, R. Blanchard, P. Genevet, and F. Capasso, “Nanometre optical coatings based on strong interference effects in highly absorbing media,” Nat. Mater. 12(1), 20–24 (2012). [CrossRef] [PubMed]
24. H. A. Macleod, Thin-film Optical Filters, 3rd ed. (Institute of Physics Publishing, London 2001).
25. H. Kim and S. M. Rossnagel, “Growth kinetics and initial stage growth during plasma-enhanced Ti atomic layer deposition,” J. Vac. Sci. Technol. A 20(3), 802–808 (2002). [CrossRef]
26. D. H. Park, C. H. Lee, and W. N. Herman, “Analysis of multiple reflection effects in reflective measurements of electro-optic coefficients of poled polymers in multilayer structures,” Opt. Express 14(19), 8866–8884 (2006). [CrossRef] [PubMed]
27. S. M. Rytov, “Electromagnetic properties of a finely stratified medium,” Sov. Phys. JETP 2, 466–475 (1956).
28. X. Wu, F. Lai, L. Lin, J. Lv, B. Zhuang, Q. Yan, and Z. Huang, “Optical inhomogeneity of ZnS films deposited by thermal evaporation,” Appl. Surf. Sci. 254(20), 6455–6460 (2008). [CrossRef]
29. K. B. Nguyen and T. D. Nguyen, “Defect coverage profile and propagation of roughness of sputter‐deposited Mo/Si multilayer coating for extreme ultraviolet projection lithography,” J. Vac. Sci. Technol. B 11(6), 2964–2970 (1993). [CrossRef]
30. A. L. Barabási and H. E. Stanley, Fractal concepts in surface growth (Cambridge University Press, 1995)
31. M. Matsubara and I. Hirabayashi, “Preparation of ultra-flat YBCO thin films by MOCVD layer-by-layer deposition,” Appl. Surf. Sci. 82, 494–500 (1994). [CrossRef]
32. R. L. Puurunen, “Surface chemistry of atomic layer deposition: A case study for the trimethyl-aluminum/water process,” J. Appl. Phys. 97(12), 121301 (2005). [CrossRef]
34. I. Perez, E. Robertson, P. Banerjee, L. Henn-Lecordier, S. J. Son, S. B. Lee, and G. W. Rubloff, “TEM-Based Metrology for HfO2 Layers and Nanotubes Formed in Anodic Aluminum Oxide Nanopore Structures,” Small 4(8), 1223–1232 (2008). [CrossRef] [PubMed]
35. P. Banerjee, I. Perez, L. Henn-Lecordier, S. B. Lee, and G. W. Rubloff, “Nanotubular metal-insulator-metal capacitor arrays for energy storage,” Nat. Nanotechnol. 4(5), 292–296 (2009). [CrossRef] [PubMed]
36. E. Hecht, “Principles of Optics”, 2nd ed. (Addison-Wesley, Reading, MA, 1987), p. 375.
37. L. Ding, T. P. Chen, Y. Liu, C. Y. Ng, and S. Fung, “Optical properties of silicon nanocrystals embedded in a SiO2 matrix,” Phys. Rev. B 72(12), 125419 (2005). [CrossRef]
38. E. S. M. Goh, T. P. Chen, C. Q. Sun, L. Ding, and Y. Liu, “Design of a Near-Perfect Anti Reflective Layer for Si Photodetectors Based on a SiO2 Film Embedded with Si Nanocrystals,” Jpn. J. Appl. Phys. 48(6), 060206 (2009). [CrossRef]