Self-assembled antireflection coatings for light trapping based on SiGe random metasurfaces

Mohammed Bouabdellaoui, Simona Checcucci, Thomas Wood, Meher Naffouti, Robert Paria Sena, Kailang Liu, Carmen M. Ruiz, David Duche, Judikael le Rouzo, Ludovic Escoubas, Gerard Berginc, Nicolas Bonod, Mimoun Zazoui, Luc Favre, Leo Metayer, Antoine Ronda, Isabelle Berbezier, David Grosso, Massimo Gurioli, and Marco Abbarchi

Phys. Rev. Materials 2, 035203 – Published 28 March 2018

Abstract

We demonstrate a simple self-assembly method based on solid state dewetting of ultrathin silicon films and germanium deposition for the fabrication of efficient antireflection coatings on silicon for light trapping. We fabricate SiGe islands with a high surface density, randomly positioned and broadly varied in size. This allows one to reduce the reflectance to low values in a broad spectral range (from 500 nm to 2500 nm) and a broad angle (up to $55^{\circ})$ and to trap within the wafer a large portion of the impinging light $(\sim 40 %)$ also below the band gap, where the Si substrate is nonabsorbing. Theoretical simulations agree with the experimental results, showing that the efficient light coupling into the substrate is mediated by Mie resonances formed within the SiGe islands. This lithography-free method can be implemented on arbitrarily thick or thin ${SiO}_{2}$ layers and its duration only depends on the sample thickness and on the annealing temperature.

Received 30 September 2017

DOI:https://doi.org/10.1103/PhysRevMaterials.2.035203

Physics Subject Headings (PhySH)

Nanophotonics

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Mohammed Bouabdellaoui^1,2, Simona Checcucci^1,3,4,*, Thomas Wood^1,†, Meher Naffouti^1,5,‡, Robert Paria Sena¹, Kailang Liu¹, Carmen M. Ruiz¹, David Duche¹, Judikael le Rouzo¹, Ludovic Escoubas¹, Gerard Berginc⁶, Nicolas Bonod⁷, Mimoun Zazoui², Luc Favre¹, Leo Metayer¹, Antoine Ronda¹, Isabelle Berbezier¹, David Grosso¹, Massimo Gurioli¹, and Marco Abbarchi^1,§

¹CNRS, Aix-Marseille Université, Centrale Marseille, IM2NP, UMR 7334, Campus de St. Jérôme, 13397 Marseille, France
²Laboratory of Physics of Condensed Matter and Renewable Energy, Faculty of Sciences and Technology, Hassan II University of Casablanca, 146 Mohammedia, Morocco
³European Laboratory for Nonlinear Spectroscopy (LENS), 50019 Sesto Fiorentino, Italy
⁴Dipartimento di Fisica ed Astronomia, Universitá degli Studidi Firenze, 50019 Sesto Fiorentino, Italy
⁵Laboratoire de Micro-optoélectronique et Nanostructures Faculté des Sciences de Monastir Université de Monastir, 5019 Monastir, Tunisia
⁶Thales Optronics, Elancourt, France
⁷Aix-Marseille Université, CNRS, Centrale Marseille, Institut Fresnel, Marseille, France

^*Corresponding author: checcucci@lens.unifi.it
^†Present address: Institut de Nanotechnologies de Lyon, Ecole Centrale de Lyon, CNRS (UMR 5270), 69134 Ecully, France; corresponding author: thomas.wood@ec-lyon.fr
^‡Present address: GREMAN LAB/ST Microelectronics 16 Rue Pierre et Marie Curie 37071 Tours, France.
^§Corresponding author: marco.abbarchi@im2np.fr

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 2, Iss. 3 — March 2018

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
Sample fabrication and particles' size distribution. (a) Diagram of the UT-SOI sample composition (12 nm of monocrystalline Si atop 20 nm of buried oxide, BOX) and description of the fabrication steps (i)–(iv). The bottom part describes the annealing cycle in the ultrahigh vacuum (UHV) of the molecular beam reactor and the Ge deposition steps. Additional details of the fabrication process are provided in the Supplemental Material [65]. (b) Statistical distribution of particle diameter for samples I (top panel), II (central panel), and III (bottom panel). The diameter of the islands is determined as $〈 L 〉 = (L_{M} + L_{m}) / 2$ , where $L_{M}$ and $L_{m}$ are respectively the longer and shorter axis of the ellipse that better approximates the particles' shape (see also Fig. SM.1 in the Supplemental Material [65]). The insets show high-resolution SEM images of the dewetted samples. For sample II the inset shows also the two-dimensional Fourier transform of a $9 \times 9 μ m^{2}$ SEM image.
Reuse & Permissions
Figure 2
Total reflectance $R_{tot}$ for Si- and SiGe-based metasurfaces. (a) $R_{tot}$ in the full investigated spectral range for a Si wafer. Experimental data, black circles; theoretical simulation for the top face only, thick purple line; theoretical simulation for both top and bottom faces, thin, orange line. The shaded area highlights the transition from above- to below band-gap frequencies. (b) Experimental data for $R_{tot}$ measured on Si wafer (gray line and shaded area), bare Si islands (empty black squares), and bare SiGe islands sample I, II, and III (respectively green circles, red squares, and blue triangles). (c) Total transmission $T_{tot}$ (blue squares) and total transmission plus total reflection $T_{tot} + R_{tot}$ (black line) for samples III. The error in the measured light intensity is less than $2 %$ at longer wavelengths where the white lamp used for illumination is less intense.
Reuse & Permissions
Figure 3
FDTD simulations. (a) Scattering intensity of an individual island having a base size of 450 nm (black line) and 150 nm red line (vertical aspect ratio $1 / 2)$ . $α, β, γ$ and $δ$ highlight the position of the main resonances where the near field intensities are evaluated (see (b)–(e)). (b) Near-field maps of the intensity of the electric field $(| E |)$ and magnetic field $(| H |)$ at 650 nm $(α)$ for the small particle simulated in (a). The white lines highlight the shape of the island and the BOX. (c)-(e) Same as (b) for the large particle: near-field maps of the intensity of the electric field and magnetic field at 750 nm $(β)$ , 960 nm $(γ)$ and 1550 nm $(δ)$ . (f) Comparison between the measured $R_{tot}$ for samples I, II, and III and the FDTD simulation taking into account a random distribution of small (150 nm in diameter) and large (450 nm in diameter) SiGe islands. Additional details of the island spatial distribution are provided in the Supplemental Material [65].
Reuse & Permissions
Figure 4
Below band-gap antireflection coating: samples I-A, I-B, and I-C and III-A, III-B, and III-C. (a) From the top to the bottom panel: total reflectance $R_{tot}$ at quasinormal incidence for samples I-A and III-A, I-B and III-B, and I-C and III-C. (b) From the top to the bottom panel: total transmission $T_{tot}$ (blue squares) and $T_{tot} + R_{tot}$ (black line) at quasinormal incidence for samples III-A, III-B, and III-C. The error in the measured light intensity is less than $2 %$ at longer wavelengths where the white lamp used for illumination is less intense.
Reuse & Permissions
Figure 5
Above band-gap antireflection coating: samples II-A, II-B, and II-C. (a) From the top to the bottom panel: total reflectance $R_{tot}$ at quasinormal incidence for samples II-A, II-B, and II-C. (b) Top panel: specular reflectance $S R$ for $s$ - (open symbols) and $p$ -polarized (full symbols) incident beam for sample II-B (triangles) and Si wafer (squares). $S R$ for a beam incident on the sample surface at angles larger than $45^{\circ}$ is measured with an ellipsometer. $S R$ at $11^{\circ}$ (shaded) is obtained from $R_{tot}$ in graph (a) and it thus also contains the reflected scattering $(R_{tot} = S R + R D$ ; see also Fig. 4). Thus it represents an overestimation of $S R$ . Bottom panel: Specular reflectance $S R$ at 900 nm for sample II-B as a function of the incidence angle normalized to $S R$ of Si wafer at the same wavelength. The error in the measured light intensity is less than $1 %$ .
Reuse & Permissions
Figure 6
Comparison between reflected scattering $(R D)$ and specular reflectance $(S R)$ for sample II-B. (a) Reflected scattering $R D$ (black triangles) and $R_{tot}$ (violet line) for sample II-B at quasinormal incidence and integrated over the half solid angle atop the sample. (b) Sketch of the experimental setup (diffusometer) used for detecting the bidirectional reflectance distribution function (BRDF). The angle of incidence of the excitation $(θ_{exc})$ and that of the detection $(θ_{\det})$ can be scanned independently over the half solid angle on the same meridian. (c) BRDF integrated from 400 nm to 800 nm for sample II-B measured for $θ_{exc}$ from $0^{\circ}$ to $- 70^{\circ}$ with respect to the normal to the sample surface and $θ_{\det}$ moving from $- 90$ to $+ 90$ . The sharp and intense peak is ascribed to $S R$ , whereas the low-intensity pedestal to $R D$ . Inset: zoom of the broad pedestal ascribed to $R D$ . The black arrow highlights a slight increase of the $R D$ for incident beam at large angles. (d) Full data set of the spectrally resolved BRDF for sample II-B for $θ_{exc} = 0^{\circ}, + 45^{\circ}$ , and $+ 75^{\circ}$ (respectively from the top to the bottom panel). Each panel reports the intensity of the BRDF as a color scale as a function of wavelength and of $θ_{\det}$ . The red dashed lines highlight $θ_{exc}$ , whereas the white dashed lines highlight the $θ_{\det}$ represented in (e). The black shaded areas hide the specular reflection. (e) Spectrally resolved BRDF for sample II-B at $θ_{exc} = 0^{\circ}, + 45^{\circ}$ , and $+ 75^{\circ}$ (respectively from the top to the bottom panel) for $θ_{\det} = - 70^{\circ}, + 10^{\circ}, + 70^{\circ}$ (top panel), $- 40^{\circ}, + 10^{\circ}, + 40^{\circ}$ (central panel), and $- 70^{\circ}, + 10^{\circ}, + 70^{\circ}$ (bottom panel). On each panel the inset displays the excitation (red arrow) and collection (brown arrows) geometry.
Reuse & Permissions

Physical Review Materials