Cathodoluminescence Imaging for Nanostructured GaN LED Materials

Part of the family of III/V semiconductors, gallium nitride (GaN) is a versatile wide-bandgap semiconductor material. It is a hard, rugged material possessing excellent electrical and optical properties. As such, it is used in a broad range of applications including LEDs for displays and lighting, high-performance electronics (high voltage, frequency, and/or temperature), and laser diodes.

For LED applications, GaN is usually grown in a heterostructure form with a number of InxGa1-xN quantum wells (QWs) which transfer the light emission from the native band-gap energy of 3.4 eV (365 nm) in the ultraviolet to the visible part of the spectrum.

Such a device is shown in cross section in Figure 1(a). It is composed of an AlGaN barrier layer, an n-doped GaN layer, p-doped GaN top layer, and 10 InGaN QWs on a sapphire substrate. The device was grown by metal-organic chemical vapor deposition (MOCVD), which is a conventional technique for growing high-quality III/V materials [1, 2].

(a) HAADF STEM image taken on a cross-sectional lamella of a (In)GaN heterostructure prepared with focused ion beam milling. The different layers are indicated in the image. For reference, the incident direction of the electron beam is also shown (b) CL spectra collected at different acceleration voltages, showing the characteristic YB, QW and GaN band gap emission (t = 250–500 ms, HV = 6–30 kV, I = 10 pA). (c) Angular CL profile (t=10 s,

Figure 1. (a) HAADF STEM image taken on a cross-sectional lamella of a (In)GaN heterostructure prepared with focused ion beam milling. The different layers are indicated in the image. For reference, the incident direction of the electron beam is also shown (b) CL spectra collected at different acceleration voltages, showing the characteristic YB, QW and GaN band gap emission (t = 250–500 ms, HV = 6–30 kV, I = 10 pA). (c) Angular CL profile (t=10 s, HV=10 kV, I=20 pA). At this voltage the emission is dominated by the QWs. (d) Cross section through (c) showing the Lambertian nature of the QW emission. Sample courtesy of Michael Latzel, Martin Heilmann, and Silke Christiansen (Max Planck Institute for the Science of Light, Erlangen). Data taken from Ref. [1]

Cathodoluminescence Imaging

Cathodoluminescence (CL) imaging and spectroscopy is widely used in the context of GaN materials and devices. The energetic electron beam can effectively excite the GaN’s wide bandgap (3.4 eV) with sub-wavelength spatial resolution, and is used for imaging local defects, for example, dislocations and probe light emission characteristics on small length scales.

Moreover, different depths can be probed by changing the energy of the primary electrons, thus enabling you to obtain more information in the case of stratified devices, as demonstrated in Figure 1(a). Using Casino Monte Carlo simulations, for instance, the electron interaction volume can be accurately estimated [3].

Spectra measured for varied primary electron energies are shown in Figure 1(b). The acceleration voltage (HV), beam current (I), and dwell time (t) are listed in the caption. For low energy (6–15 keV), the blue QW emission at 2.77 eV (448 nm) dominates the spectrum since the electron beam interaction volume remains close to the QWs.

For higher energies (15–30 keV), the electrons penetrate deep enough to excite the n-doped GaN layer beneath the QWs, resulting in two additional peaks in the CL spectrum: the characteristic “yellow band” defect luminescence peak at 2.14 eV (579 nm) related to defects in the GaN material and a peak in the UV corresponding to the band edge emission of GaN at 3.38 eV (367 nm).

The FWHM, peak center positions, and amplitudes can be accurately measured at small-length scales by using CL spectroscopy. This spectroscopic information can be used for failure/defect analysis, growth optimization, and verifying the (local) homogeneity in the optical response, for example.

In addition, the angular profile of emission which offers another significant metric for light-emitting devices can be established through detailed characterization of the spectra, angle-resolved CL imaging. Figure 1(c, d) demonstrates that the emission has a Lambertian profile in this case, which provides an insight into the carrier recombination direction in the QWs [1]. Polarization-resolved CL could also be helpful in this case.

Miniaturized LED Devices

Even though, the above instance presents a bulk LED material, there is a huge interest in miniaturizing these LED devices. This has various benefits:

  1. There is a need for nanoscale light-sources for sensing, data transmission, and transduction in nanophotonic chips and devices.
  2. The light outcoupling can be enhanced in a nanostructured LED compared to a bulk system, consequently boosting the efficiency of the device.
  3. Waveguiding and resonant effects can be employed to customize the emission polarization, angular profile, and emission spectrum to produce a more optimized and flexible device.

Due to the small feature sizes, such types of devices are hard to study with traditional optical methods. Electrical characterization is also difficult as the electrical contacting is time consuming and more complex, specifically for individual structures as that usually involves probe contacting or lithographic methods.

CL spectroscopy offers a contactless method of studying the properties of such structures. Figure 2(a) illustrates SEM images of microrod LED structures comprising of a GaN core covered with three radial InGaN QWs in a heterostructure geometry, grown using epitaxial MOCVD on a sapphire substrate [4,5].

The emission for electron excitation of the top facet is mapped with the help of hyperspectral CL imaging (see Figure 2(b)). In this image, each pixel contains a CL spectrum of which a selection is demonstrated in Figure 2(c). There is a major variation in the spectra on length scales less than 100 nm indicating the resolving power of CL imaging.

(a). Top view SEM image of a GaN nanorod. On the left, a similar nanorod is shown lying on its side where the white box indicates the part under investigation. (b) False color RGB CL image derived from a hyperspectral CL dataset where the CL intensity is divided in three color bins in the spectral range from 390 to 470 nm (t = 100 ms , HV = 5 kV, I = 120 pA). The differences in color indicate changes in the emission spectrum. (c) Representative CL spectra of different areas as indicated in (b). Inset: the CL intensity distributions at 395 and 460 nm (12 nm bandwidth), extracted from a hyperspectral image acquired on the horizontal wire shown in (a). For reference the geometrical outline of the rods is indicated. (d) Angular profiles for excitation at the same positions. Patterns were measured for 400 nm center wavelength using a 40 nm band pass color filter (t = 10 s, HV = 5 kV, I = 120 pA). Sample courtesy of Christian Tessarek, Martin Heilmann, and Silke Christiansen (Max Planck Institute for the Science of Light, Erlangen).

Figure 2. (a). Top view SEM image of a GaN nanorod. On the left, a similar nanorod is shown lying on its side where the white box indicates the part under investigation. (b) False color RGB CL image derived from a hyperspectral CL dataset where the CL intensity is divided in three color bins in the spectral range from 390 to 470 nm (t = 100 ms , HV = 5 kV, I = 120 pA). The differences in color indicate changes in the emission spectrum. (c) Representative CL spectra of different areas as indicated in (b). Inset: the CL intensity distributions at 395 and 460 nm (12 nm bandwidth), extracted from a hyperspectral image acquired on the horizontal wire shown in (a). For reference the geometrical outline of the rods is indicated. (d) Angular profiles for excitation at the same positions. Patterns were measured for 400 nm center wavelength using a 40 nm band pass color filter (t = 10 s, HV = 5 kV, I = 120 pA). Sample courtesy of Christian Tessarek, Martin Heilmann, and Silke Christiansen (Max Planck Institute for the Science of Light, Erlangen).

These changes can be related to the rod’s growth and internal structure. The emission bands seen around 460 nm (2.7 eV) and 395 nm (3.13 eV) at positions 3, 5, and 6 are strongest at the edge. This emission is also localized at the top of the rod as can be observed in the inset in (c) where CL intensity maps are illustrated, taken on a horizontal rod (see Figure 2(a)).

By correlating the CL data with TEM data (not depicted in this article), these emission bands are attributed to r-plane and m-plane (specific crystal planes in the wurtzite GaN crystal) InGaN QWs, respectively. The protrusion on the top visible in the SEM and CL images arises from the fact that this part of the rod has a varied polarity from the surrounding material [4]. It depicts emission bands at 420 nm (3.0 eV) and 400 nm (3.1 eV) for positions 1 and 2 respectively, different from the QW emission at the edge.

In addition to the variations in the emission spectra, there are also strong differences in the angular profile. Angular profiles are depicted in Figure 2(d) for the positions signified in (b). The patterns are relatively different from the bulk material and also rely strongly on the electron beam’s impact position. The observed angular pattern is affected both by the internal material structure and the (local) geometry. Therefore, it can be used to get a better understanding of the (optical) properties [4-7]. In addition, designs that are improved for a specific directivity can be quantitatively tested.

While this article illustrates separate CL images for a broken rod lying down and for the top-facet, a more complete bird’s eye view of a single structure can be achieved by mounting the sample on a tilted holder, thus enabling simultaneous CL imaging of side and top facets. The type of CL imaging shown in this article can certainly be extended to other III/V compounds like GaAs and InP, which are pertinent for integrated photonics, photovoltaic applications, and IR light sources among others [6, 7].

References

[1] S. Meuret et al. Phys. Rev. B 98, 035308 (2017).

[2] M. Latzel et al. Nanotechnology 28, 055201 (2017).

[3] H. Demers et al. Scanning 33, 135-146, (2011).

[4] C. Tessarek et al. J. Appl. Phys 114, 144304 (2013).

[5] C. Tessarek et al. submitted (2017).

[6] B. J. M. Brenny et al. Appl. Phys. Lett. 107 201110 (2015).

[7] B. J. M. Brenny et al. ACS Photon. 3, 677 (2016).

This information has been sourced, reviewed and adapted from materials provided by Delmic B.V.

For more information on this source, please visit Delmic B.V.

Citations

Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Delmic B.V.. (2018, December 18). Cathodoluminescence Imaging for Nanostructured GaN LED Materials. AZoNano. Retrieved on June 24, 2019 from https://www.azonano.com/article.aspx?ArticleID=5072.

  • MLA

    Delmic B.V.. "Cathodoluminescence Imaging for Nanostructured GaN LED Materials". AZoNano. 24 June 2019. <https://www.azonano.com/article.aspx?ArticleID=5072>.

  • Chicago

    Delmic B.V.. "Cathodoluminescence Imaging for Nanostructured GaN LED Materials". AZoNano. https://www.azonano.com/article.aspx?ArticleID=5072. (accessed June 24, 2019).

  • Harvard

    Delmic B.V.. 2018. Cathodoluminescence Imaging for Nanostructured GaN LED Materials. AZoNano, viewed 24 June 2019, https://www.azonano.com/article.aspx?ArticleID=5072.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this article?

Leave your feedback
Submit