Abstract

Semiconductor multilayer and device fabrication is a complex task in electronics and opto-electronics. Layer dry etching is one of the process steps to achieve a specific lateral device design. In situ and real-time monitoring of etch depth will be necessary if high precision in etch depth is required. Nondestructive optical techniques are the methods of choice. Reflectance anisotropy spectroscopy equipment has been used to monitor the accurate etch depth during reactive ion etching of III/V semiconductor samples in situ and real time. For this purpose, temporal Fabry–Perot oscillations due to the etch-related shrinking thickness of the uppermost layer have been exploited. Earlier, we have already reported an etch-depth resolution of ±16.0 nm. By the use of a quadruple-Vernier-scale measurement and an evaluation protocol, now we even improve the in situ real-time etch-depth resolution by a factor of 20, i.e., nominally down to ±0.8 nm.

Highlights

  • Reactive ion etching (RIE) is widely applied in crystalline semiconductor technology, e.g., for the etching of multilayered III/V semiconductor samples like GaAs/AlxGa1−xAs heterostructures.[1–8] Usually, the etch process has to be stopped at a certain depth in order to meet the desired device design

  • In Ref. 9, we have shown that an etch-depth monitoring resolution of ±16.0 nm can be achieved using Fabry–Perot oscillations at one single photon energy

  • Depending on exact conditions, sometimes the genuine Reflectance anisotropy spectroscopy (RAS) signal according to Eq (1) shows Fabry–Perot oscillations with strong modulation/contrast, and sometimes this is rather true for mean reflectivity ⟨R⟩ [the denominator in Eq (1)]

Read more

Summary

INTRODUCTION

Reactive ion etching (RIE) is widely applied in crystalline semiconductor technology, e.g., for the etching of multilayered III/V semiconductor samples like GaAs/AlxGa1−xAs heterostructures.[1–8] Usually, the etch process has to be stopped at a certain depth in order to meet the desired device design. For specific devices (see, e.g., Ref. 9), this might even require an etch-depth control with accuracies better than ±20 nm. Clusters on the etch front break the symmetry and allow for reflectance differences for linearly polarized light depending on its plane of polarization. Not this basic RAS principle is used here but rather the RAS equipment with broadband (1.5–5.0 eV) light incidence perpendicular to the etch front. The occurrence of temporal Fabry– Perot oscillations (maxima = antiresonances due to the reflective setup) is exploited. Their temporal period depends on photon energy and current etch rate (typically on the order of 70 nm/min). The oscillations can be observed either in the transients of the RAS signal for specific photon energies or in the so-called RAS color plot, which shows the temporal evolution of the RAS spectra with color-coded signal heights

IMPROVING THE RESOLUTION OF ETCH-DEPTH MONITORING
DATA RETRIEVAL AND EVALUATION
FURTHER CHECK OF APPLICABILITY OF THE PROCEDURE
Findings
CONCLUSIONS
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call