Improving the performance of p-type GaN LEDs with argon

If you want to get a significant doping profile and low resistance p-type GaN, then consider replacing your carrier gas with hydrogen to argon.

Author: Vladimir Dmitriev, Alexander Usikov
Unit: Technologies and Devices International

Although the development of GaN LEDs has been quite successful, the use of organometallic chemical vapor deposition (MOCVD) processes has rendered their p-doped regions undesirably desirable. What is required for a true luminance illuminator such as a solid-state illumination target is a highly doped p-type region with low resistance and discontinuous doping profile in an epilayer stack structure, but Today's processes produce high-impedance materials that impede current spreading, which can increase operating voltages and hinder high drive currents when poor ohmic contacts occur.

This event originated from p-type GaN and AlGaN, which were developed using standard dopant-magnesium for organic metal vapor deposition (MOCVD), and these nitrides have high resistance, in part because of the growth of the coating. After the completion of the heat activation caused by. However, even if the p-type GaN thus treated has a resistance of 0.1 Ω-cm or more, the luminous efficiency of the LEDs of the highest brightness is limited due to the need for a higher operating voltage.

The cause of these high resistances may be due to the passivation of hydrogen on the surface (hydrogen is a standard MOCVD carrier gas). Thermal activation after film formation destroys the magnesium-hydrogen complex and promotes electrical conductivity, even if too much magnesium is doped to cause deterioration in material quality.

Nitrides doped with magnesium by the organometallic chemical vapor deposition (MOCVD) method are also affected by the "memory effect", which will stay in the coating chamber after the linear vacuum tube is closed, and then It is absorbed during the subsequent film formation process in which no magnesium component is present.

In response to the above problems, we have developed a thin film deposition process that does not use pure hydrogen. The hydride vapor phase epitaxy (HVPE) process technology used in this research work (sponsored by the US Department of Energy's Solid State Lighting Program) uses ammonia (NH3) and hydrogen chloride (HCl) as the reaction gas and argon as the carrier gas. Higher deposition rates and lower dislocation densities can be achieved.

Organometallic vapor phase deposition (MOCVD) crystal growth with argon as the carrier gas is not more effective because the Group III evaporation source requires a higher concentration of NH3, which causes more hydrogen to flow to the sediment layer. in.

We used a patented multiwafer HVPE reactor to deposit a p-type GaN layer on the (0001) c-plane sapphire, and the deposition conditions were set at a typical rate of 1 μm/min. 1050 oC, and use high-purity gallium material as the source of III group evaporation, with magnesium as a dopant.

According to the secondary ion mass spectrometry (SIMS) measurement, we fabricated a GaN layer with a thickness of 3–15 μm and a concentration of 2 × 1016 cm–3 – 2 × 1020 cm–3 magnesium. All epitaxy The layers (epilayers) have a smooth surface, while the X-ray diffraction measurements show that high doping concentrations do not affect the quality of the crystal. Activation is not required in the sample to produce p-type conductivity. The results of the capacitance-voltage (CV) measurement using a mercury probe revealed p-type conductivity at a magnesium atom concentration of 1 × 1017 cm–3, and the CV test results also showed a net receptor concentration (receptor). The number minus the number of donors) is as high as 3 × 1019 cm–3, and such doping concentration is just right for forming a good ohmic contact.

The electrical measurement results of the deposited layer with a carrier concentration range of 4 × 1017 – 1.5 × 1018 cm–3 show that the resistivity is 0.02–1.00 Ω-cm, and the hole mobility can at least reach a film like MOCVD-deposited. The electrical results are equally good, and the results indicate that the LEDs deposited by HVPE should have fairly good current-spreading characteristics. The results of the secondary mass spectrometer (SIMS) measurement also show that our materials are not susceptible to the memory effect (Figure 1).

Higher concentrations of magnesium than hydrogen (Figure 2) cause high conductivity and high net acceptor concentration. According to our measurements of two (1–3) × 1019 cm–3 magnesium atom concentration doping samples, Minimizing hydrogen concentration is critical for high doping. The sample produced by the hydrogen concentration of 4 × 1017 cm–3 has a net acceptor concentration of 1.2 × 1019 cm–3, while the net concentration of the film produced by the hydrogen concentration of 1 × 1018 cm–3 is only 1.3 × 1018 cm–3. The change in hydrogen concentration causes almost all of the magnesium in one sample to be activated and only 10% in the other parts to be activated.

We have verified the diversity of our luminescent HVPE film deposition methods. "Upside-down" LEDs that cannot be fabricated by organometallic chemical vapor deposition (MOCVD) can also be fabricated on sapphire-based p-type GaN templates. The InGaN layer in these template structures contains 15–30 mole % of InN. The composition has an emission wavelength between 450 and 515 nm.

Figure 1: The secondary ion mass spectrometer (SIMS) of the second, third and fifth layers deliberately doped with magnesium atoms in a multilayer structure shows discontinuous doping structure and a little magnesium The atoms were inadvertently doped into the fourth layer, and the results on the surface of the sample are shown in the figure.

Figure 2: According to the results of secondary ion mass spectrometry (SIMS), the doping concentration of magnesium is higher than the background concentration of hydrogen, wherein the dotted line represents the equivalent concentration of magnesium and hydrogen atoms.

Author
Vladimir Dmitriev is the chairman and CEO of TDI, who died on January 6, 2008.
Alexander Usikov is the head of R&D and a senior scientist at TDI. His current research focuses on III-N semiconductor physics and novel technologies. The authors are grateful to the Army Research Laboratory and the Palo Alto Research Center for their assistance in LED manufacturing, and to the US Department of Energy, the Department of Commerce and the Department of Defense for funding, and TDI engineers. And the research and contributions of the team of scientists, and the contributions of Oleg Kovalenkov, Vitaly Soukhoveev and Vladimir Ivantsov in this study.

SMD LED

Wuxi Ark Technology Electronic Co.,Ltd. , https://www.arkledcn.com