In Situ Etch Rate Measurements by Alpha-Particle Energy Loss
- PDF / 859,419 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 20 Downloads / 176 Views
application of the technique we will describe measurements of the temperature dependence of the etch rate of GaAs in the 15-150 °C temperature range using optical bandgap thermometry to determine the substrate temperature. In a second example, we explore the application of the technique to etch rate of short pitch (250-500nm) grating. In this case the shape of the alphaspectrum is sensitive to the profile of the etched trenches. INTRODUCTION Due to the need for high quality, reproducibility and flexibility in the fabrication of advanced semiconductor devices there is an increasing use of plasma etch systems, such as electron cyclotron resonance (ECR). These systems can provide anisotropic smooth patterning at a large range of etch rates [1, 2]. The increased performance and the complexity of semiconductor devices have created a need for improvement in the control of the etch processes. In many cases an accuracy of 1% or better of the etch depths is required. A number of techniques have been explored for accurate measurements of the etch rates. These include ellipsometry, thin film interferometry, mass spectrometry and optical emission spectroscopy [3-61. The last two measure the residual gases during the etch and not the layer itself, and may not be accurate enough at low etch rates. The optical monitoring techniques are sensitive to the optical properties of the etched material which, in general, depend in a complex way on the thickness and on the composition of 29 Mat. Res. Soc. Symp. Proc. Vol. 569 01999 Materials Research Society
the etched layer. These techniques can reach high accuracy with some restrictions - large area monitoring and layers with optical contrast. They cannot be applied, for example, to monitoring of homoetching. In this paper we present the first in situ implementation of a different approach to etch rate measurements, namely the alpha-particle energy loss method. This technique has been used by us for real-time monitoring of film thickness and composition during growth by molecular beam epitaxy (MBE) [7]. The method, described in more detail in previous publications [8,9], measures directly the absolute thickness of the layer and does not depend on optical properties or on fitting data to complex models. Prior to growth, the bare substrates are placed facing a radioactive 218Th source inside a vacuum chamber. During the decay of this isotope, the generated isotope of 224Ra is being recoil implanted into the substrates. The alpha-emitting 224Ra decays with a half-life of 3.66 days through a decay chain emitting alpha-particles of several characteristic energies [10]. As the alpha-particles pass through a layer of thickness X, they lose an average amount of energy,
E1,, =X(dE/ld)
(1)
where dE/dX is the characteristic energy loss per unit length for the material of the film. Since dE/dX values are known for most elements and they are insensitive to chemical bonding configurations for MeV-energy alpha-particles, a measurement of Eo,, 0 gives directly the thickness X for films of known com
Data Loading...
