Limitations of Cl₂/O₂‐based ICP‐RIE of deep holes for planar photonic crystals in InP

Abstract

A detailed study of dry‐etching of high‐aspect‐ratio holes into an indium phosphide substrate is presented for a Cl2/O2‐based plasma chemistry. The etching is performed in an inductively coupled plasma reactive ion etching reactor. The separate influence of the various etching parameters on the quality of the etched holes is identified. Quality measures such as high aspect ratio, hole cylindricity and verticality as well as sidewall smoothness can be controlled by varying the ICP power, the relative O2 flow rate and the self‐bias of the plasma. We were able to clearly identify trade‐offs that have to be made and limitations of the etching chemistry/technology used: If the aspect ratio improves, then the cylindricity also improves, whereas the verticality and the sidewall smoothness degrade. In previous reports, a certain ambiguity is generally observed in the sense that different process parameters exhibit partially contradicting trade‐offs. We show that this behaviour can be remedied by a careful selection of the variable parameters. (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Self-aligned mask renewal for anisotropically etched circular micro- and nanostructures

“…The fabrication of such holes in InP typically relies on a hardmask of SiO 2 [11] or SiN x [2], which offers a good selectivity to the substrate. Standard etching techniques include chemically assisted ion beam etching (CAIBE) [12] and ICP-RIE [2,5]. In spite of a high selectivity between SiN x and InP (∼15-25, depending on the etching chemistry), it is nevertheless the mask thickness which sets the main limit to the maximum hole depth that can be reached.…”

“…(3) Modifications to the etching chemistry: the selectivity between mask and substrate material is often strongly dependent on the chemical composition of the etching plasma. If the limitations of a given chemistry are reached [5], then the assessment of alternatives can be considered. This is often done empirically and requires considerable effort.…”