Menu
Cart 0

What is Molecular-Beam Epitaxy (MBE)?

This is a continuation from the previous tutorial - what are vapor-phase epitaxy (VPE) and metal-organic vapor-phase epitaxy (MOVPE)?

 

In the molecular-beam epitaxial (MBE) technique, epitaxial layers are grown by impinging atomic or molecular beams on a heated substrate kept in an ultrahigh vacuum. The constituents of the beam "stick" to the substrate, resulting in a lattice-matched layer.

The beam intensities can be separately controlled to take into consideration the difference in sticking coefficients of the various constituents of the epitaxial layers.

The widespread use of MBE for the growth of different III-V semiconductors resulted from the original work of Arthur and Cho. The lattice-matched growth of AlGaAs on GaAs substrates by MBE was first reported in 1971.

Since then, extensive work on heterostructure lasers, microwave devices, quantum-well lasers, and superlattice structures has been reported using AlGaAs materials prepared by MBE.

Figure 4-12 shows a schematic of an MBE system for the growth of AlGaAs heterostructures.

 

Figure 4-12. Schematic illustration of a molecular-beam-epitaxy (MBE) system for epitaxial growth of AlGaAs.

 

The substrate is heated by a Mo heating block inside a vacuum chamber at a pressure of \(10^{-7}-10^{-10}\) torr. The electron-diffraction gun, the Auger analyzer, and the mass spectrometer can be used to study the layer characteristics during growth.

The sources Ga, GaAs, and so forth are kept in independently heated effusion ovens enclosed in liquid-nitrogen-cooled shrouds. The dopants also have separate effusion ovens.

The effusion ovens are heated to a temperature high enough to produce an adequate beam flux on the surface. For the growth of GaAs on AlGaAs, the Ga flux is in the range of \(10^{12}-10^{14}\text{ atoms/(cm}^2\cdot\text{sec)}\).

The flux needed depends on the material's sticking coefficient (the fraction of atoms in the beam sticking to the substrate). For many group III elements (e.g., Al or Ga) the sticking coefficient is nearly unity.

However, the coefficient differs significantly for group V elements. This makes it necessary to determine empirically the sticking coefficient for all the components as a function of substrate temperature and beam intensity.

Hsang et al. reported on InGaAsP (\(\lambda=1.3\) μm) double-heterostructure lasers fabricated by MBE. Separate In and Ga beams were used, and the substrate temperature was approximately \(580-600^\circ\text{C}\).

To maintain the correct As-to-P ratio, a single beam of As and \(\text{P}_2\) was used. It was obtained by passing \(\text{As}_4\) and \(\text{P}_4\) (evaporated from elemental arsenic and red phosphorus) through a common high-temperature zone.

Panish and coworkers have demonstrated the growth of high-quality InGaAsP layers using gas sources for \(\text{As}_2\) and \(\text{P}_2\) beams. The use of gas sources for the growth of GaAs by MBE has also been reported.

The \(\text{As}_2\) and \(\text{P}_2\) molecules are generated by decomposing \(\text{AsH}_3\) and \(\text{PH}_3\) in a heated chamber and then allowing them to leak out to the effusion section connected to the MBE growth chamber.

Figure 4-13 shows a schematic drawing of two types of gas sources.

 

Figure 4-13. Schematic illustration of two-types of high-pressure gas sources used for MBE growth of InGaAsP.

 

The major difference between the two types is that one uses separate decomposition tubes and the other uses the same tube.

The decomposition tubes are operated at \(900-1200^\circ\text{C}\) and are filled with gases at pressures in the range of 0.3-2.0 atm. The tubes have a leaky seal at one end.

At the operating temperatures and pressures stated above, \(\text{AsH}_3\) and \(\text{PH}_3\) decompose to produce \(\text{As}_4\), \(\text{P}_4\), and \(\text{H}_2\). The molecules leak into a low-pressure (few millitorr), heated region where \(\text{As}_4\) and \(\text{P}_4\) decompose further into molecules of \(\text{As}_2\) and \(\text{P}_2\).

Panish has used a low-pressure gas source where \(\text{AsH}_3\) and \(\text{PH}_3\) decompose at a pressure of less than 0.1 torr. Calawa used similar low-pressure sources for MBE growth of high quality GaAs.

Well-controlled molecular beams of \(\text{As}_2\) and \(\text{P}_2\) can be obtained using gas sources of the type shown in Figure 4-13. The beam intensity can be easily controlled by varying the pressure of \(\text{AsH}_3\) and \(\text{PH}_3\) in the decomposition chamber.

The gas-source MBE technique has been use for the fabrication of high-quality InGaAsP double-heterostructure lasers.

Another growth technique, known as chemical-beam epitaxy (CBE), has been demonstrated by Tsang for the growth of InP and GaAs.

In this technique, all sources are gaseous and are derived from group III and group IV alkyls. Figure 4-14 shows a schematic diagram of a CBE system.

 

Figure 4-14.  Schematic illustration of a chemical-beam epitaxy (CBE) system obtained by modifying a conventional MBE apparatus (see Figure 4-12).

 

The growth chamber is similar to that of a conventional MBE system and is kept at high vacuum (\(\lt5\times10^{-4}\) torr).

In and Ga are obtained by pyrolysis of either trimethylindium (TmIn) triehylindium (TEIn), and trimethylgallium (TMGa) or triethylgallium (TEGa) at the heated substrate surface.

The \(\text{As}_2\) and \(\text{P}_2\) molecules are obtained by thermal decomposition of trimethylarsine (TMAs) and triethylphosphine (TEP) passing through a Ta- or Mo-buffered, heated alumina tube at a temperature of \(950-1200^\circ\text{C}\).

The CBE technique differs from MOVPE in the following way: in CBE the metal alkyls impinge directly on the heated surface, leaving the metal at a high substrate temperature, whereas in low-pressure MOVPE the metal alkyls along with \(\text{H}_2\) form a stagnant boundary layer (as shown in Figure 4-11) [refer to the what are vapor-phase epitaxy and metal-organic vapor-phase epitaxy tutorial] over the heated surface.

Thus growth rate of the epitaxial layer for the MOVPE is diffusion-limited, whereas for CBE it is limited by the flow rate of metal alkyls.

 

 

The next tutorial explains lattice-mismatch effects on semiconductor epitaxial growth.


Share this post


Sale

Unavailable

Sold Out