Tackling the amber emission challenge with cubic nitrides

The growth of cubic nitride by MBE promises to enable the production of microLEDs that span the entire visible range.

BY JÖRG SCHÖRMANN, MARIO ZSCHERP, SILAS JENTSCH AND SANGAM CHATTERJEE FROM JUSTUS LIEBIG UNIVERSITY GIESSEN AND VITALII LIDER, ANDREAS BEYER AND KERSTIN VOLZ FROM PHILIPPS-UNIVERSITY MARBURG

What sounded like science fiction a mere two decades ago is starting to become part of our everyday life: virtual elements enriching the real world. This is evident in the on-going trend of companies and organisations turning to immersive technologies to benefit from the advantages of extended reality, such as augmented reality (AR) or mixed reality (MR). However, while much progress has been made with these technologies, there are still a few hurdles to overcome.

One of the biggest is to address the poor visibility of displays under bright ambient conditions, such as sunlight. A promising candidate for tackling this weakness is the microLED. However, this emitter, with dimensions well below 10 µm, needs to deliver an improved performance – as well as, ideally, the emergence of a unified materials platform covering the whole visible spectrum.

Addressing the latter issue requires some radical changes to LED production. Today, efficient, large-area white and blue LEDs employ hexagonal nitride technology, while their red counterparts are based on AlGaInP multilayers with a different crystal structure. Regardless of the approach, these devices are held back by a decrease or ‘droop’ in efficiency as the emission wavelength is extended to the green/amber range. This flaw is commonly known as the ‘green gap’.

Figure 1. The quality of material can be evaluated by considering values provided by w-scans, produced by X-ray diffraction. The general trend is a decreasing value for ∆w with increasing layer thicknesses. Grey squares and circles refer to other c-GaN layers grown by MOCVD and MBE on comparable 3C-SiC substrates (left). An atomic force microscopy image offers insight into surface morphology, uncovering mono-atomic steps across the smooth surface of c-GaN (right).

Of these two classes of semiconducting materials, the group-III nitrides – they include AlN, GaN, InN and their alloys – are preferred, being viewed as the crucial semiconductor materials for the current and next-generation of optoelectronic technologies. Up until now, all relevant devices made from this material system are based on the thermodynamically stable hexagonal phase.

At the very heart of these III-nitride devices, which emit in the blue and green, are InGaN multilayer structures that incur inherent internal electric fields. This gives rise to the quantum confined Stark effect (QCSE), which result from the interplay of the spontaneous polarisation that arises from the inherent asymmetry of the hexagonal crystal structure, and the piezoelectric polarisation, an undesirable effect of strain in this crystal structure. One culmination of the QCSE, which impedes radiative recombination, is a spatial separation of the electrons and holes in the quantum wells.

These issues are particularly prevalent at longer wavelengths. To propel emission from the blue to the green and beyond, the indium content in the InGaN quantum wells has to increase. But this induces even more strain in the active region, enhancing the QCSE and driving down the radiative recombination rate.

Figure 2. TX-ray diffraction of the asymmetric (113) reflection of c-GaN. The reciprocal space map reveals a pure cubic phase and a single, partially strained c-In_0.47Ga_0.53N layer.

Another side effect of the increase in indium richness in the wells is a screening of the QCSE under high injection conditions. A decrease in colour stability results – although this can be turned into an advantage, if there is a need for colour tuning.

For phosphide LEDs there is a different set of issues at play. When miniaturising these devices, surface recombination drags down efficiency, due to an increased surface-to-volume ratio. While it might be possible to address this concern, there is also a more fundamental one that cannot be overcome. With this material system, as the composition is adjusted to shift the emission towards the green, the bandgap switches from direct to indirect. This causes efficiency to plummet, and prevents the use of phosphides for the manufacture of blue and green LEDs.

The cubic advantage
Intriguingly, group III-nitrides can also crystallise in the metastable cubic structure. This form promises several advantages, stemming from a polarisation-free material system that is inherent, thanks to cubic crystal symmetry. The most prominent merit of such a structure is that it is free from the QCSE, equipping LEDs with superior colour stability at higher injection currents. Additional advantages of c-GaN over its hexagonal counterpart include a high carrier mobility, a high p-type conductivity and a high electron drift velocity.

As well as all these attributes, c-GaN has a bandgap of 3.2 eV. That’s 0.2 eV lower than h-GaN, reducing the necessary indium content by about 10 percent for targeting a particular emission wavelength in the red.

Due to the combination of an absence of the QCSE and a lower bandgap, cubic nitrides promise to provide a material platform for LEDs spanning the entire visible range. The opportunity to produce powerful blue, green, and red emitters from this material system is incredibly attractive for cost-efficient integration schemes, in particular, once ported onto 300 mm silicon wafer technology.