By: W. J. Alford, T. D. Raymond, and J. L. Reno
Molecular Beam Epitaxy (MBE) is a common technique for producing semiconductor devices. The technique uses atomic beams incident on heated substrates to grow materials such as AlGaAs.
The atomic beams are generated by heating a crucible containing the elemental source material.
The atomic beam fluxes determine the growth rate and composition of the material.
Currently, beam fluxes are measured and adjusted prior to the growth using Reflection High Energy Electron Diffraction (RHEED) — a technique that requires interruption of the growth process and, hence, cannot be used during growth. Since beam fluxes can change during growths due to source aging, an in-situ monitor of beam fluxes is desirable for precise control of growth rate and material composition.
We are developing a laser-based system for monitoring the Group III elements Al, Ga, and In.
We use atomic absorption of resonant laser light to determine the atomic beam fluxes. Single-frequency, near-infrared lasers are frequency doubled and passed within an inch of the heated substrate where growth is occurring. The image above shows a picture of the heated substrate and the atomic fluorescence which results from absorption of the laser. Since single-pass absorptions can be small, a multipass (22 passes) configuration is used to enhance the amount of absorption. We are currently using a rather large laser system (argon-ion-laser-pumped Ti:Sapphire system) but have developed compact diode laser systems for this application. We have demonstrated the ability of this Optical Flux Monitor (OFM) technique to precisely measure atomic fluxes by comparing the OFM-measured growth rate to that determined by RHEED.
The results, shown in the lower graph of Figure 1 below, indicate we can measure the atomic beam flux to a precision of better than 1.5%. Our goal is to reach a precision of 0.1%. We think this goal may be achievable with more advanced spectroscopic techniques than currently being used.
Atomic absorption allows precise determinations of the Group III fluxes but will not work well for determining much smaller dopant fluxes. Many semiconductor devices have small concentrations of dopants such as Si and Be in order to modify the electrical properties of the material. We have demonstrated that Laser Induced Fluorescence (LIF) has sufficient sensitivity to be used as a monitor of dopant fluxes during MBE growth. Convenient laser sources for the UV wavelengths required for dopant detection are not currently available and will require further development.
Optical flux monitoring promises to become a useful tool for real time measurement of atomic fluxes. When combined with a closed-loop control system, the improved control of material composition and growth rate should allow new, more challenging devices to be grown.
DP sponsors various phases of this work.
Point of Contact: Greg Hebner