Page 171 - Semiconductor Manufacturing Handbook

Geng(SMH)_CH12.qxd 04/04/2005 19:49 Page 12.20

PLASMA ETCHING

12.20 WAFER PROCESSING

12.5 ENDPOINT DETECTION IN PLASMA ETCHING

In many instances, overetch is mandatory primarily thanks to surface topography and etching uniformity
implications. Yet overetch must be done with great care. For example, in modern MOS devices the gate
oxide is extremely thin—less than 150 Å—rendering excessive overetch undesirable. Though high selec-
tivity can be used to mitigate the situation, some requirements in anisotropy often compromise the high
selectivity. Therefore, endpoint detection plays an important role in plasma etching processes.
The sequence of events in a typical plasma etching process provides many possibilities for detect-
ing the endpoint. When RF power is turned on, gases are ionized and etchant species are produced.
These species react with substrate materials to produce the primary volatile products. Toward the end
of the etch cycle, the underlying materials begin to be exposed and react with the etchant species to
produce the secondary volatile products. Meanwhile, the concentration of the primary volatile prod-
ucts decreases. In the overetching period, the primary and secondary productions reach steady states.
A variety of techniques can be constructed in principle:
• Pressure change: Pressure increases initially when gas starts to dissociate. During etching, reac-
tions consume gas molecules and thus cause a decrease in system pressure. After etching reaches
the endpoint, pressure rises again. This technique was used in early batch etchers. Yet in modern
dry etchers that maintain constant pressures, it is no longer used.
• Bias change: During an etch cycle, the impedance and sheath characteristics change. This usu-
ally causes changes in the self-bias voltage in asymmetrical reactor geometries. In photoresist
stripping, this is an effective technique for endpoint detection.
• Mass spectrometry: Different chemical species are associated with the start and end of an etch
−5
cycle, and they can be monitored by mass spectrometry. In high vacuum—less than 1 × 10 torr—
the process gas is sampled and specific etch products are detected.
• Laser interferometry and reflectance: When monochromatic light rays are reflected from the
front and underlying interface of a transparent film, constructive and destructive interference takes
place. At normal incidence, interference maxima and minima occur when twice the thickness of
film (d) is the multiple of the wavelength (l) divided by the refractive index (n), as expressed in
the following:

d = l n ⋅
2 (12.4)
In an etch cycle, the intensity of a monitoring laser beam varies sinusoidally when the film thick-
ness changes. On the other hand, metal films are opaque but usually have high reflectance. So etch-
ing through a metal-semiconductor or metal-insulator interface is marked by a sharp change in the
reflected laser intensity.
• Optical emission spectroscopy (OES): Many chemical species from etchant and etch products are
present in plasma and optical emissions can emanate at different wavelengths. An optical spec-
trometer can be set to the line of interest and follow its intensity during an etch cycle. This tech-
nique is capable of endpoint detection with great sensitivity.
In practice, laser interferometry and OES are the two most commonly used techniques in plasma
etching. They require only a suitably located optical window on the chamber and can be easily
implemented. More elaborate data processing is often required to sort through collected spectra. One
example is illustrated in a HEMT fabrication process. 101 In this case a gate was defined in 2500 Å of
GaAs on 500 Å of Al Ga As using a nitride mask. Figure 12.16a shows the optical emission
0.25 0.75
spectra collected from main etch and overetch periods in the process. It is apparent that the “straight”
OES endpoint strategy is not obvious in detecting the endpoint. However, the OES spectra can be
processed to elucidate the difference between the main etch and overetch spectra with a focus on the
403-and 417-nm Ga lines. Figure 12.16b shows the derivative of the Ga signal. A clear and distinct
endpoint was determined.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)
Copyright © 2004 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.