Semiconductor Sources: Progress is made on zinc oxide emitters

Zinc oxide materials have the potential to give GaN a run for its money as a material for visible and UV light sources. At the fall 2005 meeting of the Materials Research Society (Nov. 29-Dec. 3, 2005; Boston, MA) researchers presented new results on manipulating the bandgaps (and hence emission wavelengths) of zinc oxide (ZnO) alloys. Jianwei Dong, a member of a group at SVT Associates (Eden Prairie, MN) headed by Andrei Osinsky, described the group’s recent results from magnesium and cadmium-doped ZnO and n-ZnO/p-AlGaN (aluminum gallium nitride)-based heterostructures.¹

Advantages of ZnO

Zinc oxide is an attractive material because it luminesces through the visible and UV and is relatively easy to work with because wet-chemical etching can be used to fabricate devices. Zinc costs much less than gallium or indium, which suggests that entire systems could cost less. Zinc oxide substrates can be made with low defect densities and can be scaled up to large sizes for production. Zinc’s low refractive index makes extracting light from the material easier—which eases a longstanding problem for many types of light-emitting diode (LED).

Moreover, because the exciton binding energy of ZnO is 60 meV (compared to the binding energy of GaN, which is less than 30 meV), devices made from ZnO can operate at high temperatures more efficiently, which could be a boon for industrial or scientific applications that require either high-temperature or temperature-robust applications. Other uses include solid-state lighting and UV sensors or emitters. The work done by Osinsky’s group was sponsored by contracts from the U.S. Army, Department of Energy, and National Science Foundation.

Before ZnO achieves wide acceptance, researchers must overcome several challenges. Although progress continues, it is still difficult to make high-quality epitaxial layers, manipulate the bandgap, and create quantum wells for devices. The SVT researchers (as well as presenters from other organizations) reported progress in all three areas.

Yet another challenge is to obtain p-type doping in zinc oxide, according to Dong. While several reports at the MRS meeting reported p-doping (including one group that made a blue homojunction LED²), researchers don’t appear to have found a way to create consistently good-quality p-doped ZnO.

Adding magnesium and cadmium

The bandgap of the alloy can be altered by adding materials with larger or smaller bandgaps than ZnO. Zinc oxide has a 3.3-eV bandgap, whereas the bandgap of magnesium oxide (MgO) is 7.9 eV and that of cadmium oxide (CdO) is 2.3 eV. Adding one or the other could shift the output wavelength further into the UV or throughout the visible, respectively.

But adding too much Mg or Cd might also alter the material’s crystal structure. Zinc oxide has a wurtzite crystal structure (characterized by alternating stacking layers of tetrahedrons), which is attractive partly because it almost matches the lattice of GaN (1.7% lattice mismatch). Both MgO and CdO have attractively different bandgaps from ZnO, but they also have cubic crystal structures. Therefore the challenge is to alter the bandgap without introducing dislocations.

Dong reported using radio-frequency-plasma-assisted molecular-beam epitaxy to grow a ZnCdO alloy film (containing as much as 78% Cd), as well as a MgZnO film. The Cd film quality wasn’t as high as ZnO, but it produced bright cathodoluminescence at wavelengths throughout the visible spectrum (see figure).

High-temperature UV LED

Although p-doping is still problematic, n-doping is possible. Osinsky’s group made a three-interface LED on a sapphire substrate, consisting of layers of MgZnO, ZnO, AlGaN, and GaN. The UV LED emitted 20 µW (at about 380 nm) , and continued to operate at temperatures above 375°C. Dong noted that threading dislocations in the AlGaN stopped at the zinc layer. The next step in the development may result in a UV laser from this material.

Much of this work was done as part of a Phase I SBIR from the U.S. Army, monitored by Michael Gerhold. In late 2005, the company started a Phase II SBIR to continue developing ZnO-based light emitters for UV and blue applications. “In principle,” says Jorge Lubguban of the University of Missouri-Columbia (Columbia, MO), “ZnO should be better, brighter, and cheaper than GaN devices.”

REFERENCES

1. A.V. Osinsky et al., 2005 MRS Fall Meeting, paper EE9/FF18.

2. A. Tsukazaki et al., 2005 MRS Fall Meeting, paper DD4.2.