Measurement of electrostatic fields in Ge-doped AlGaN structures by off-axis electron holography
Authors
Alexis Palais (1), David Cooper (1), Eva Monroy (2), Bérangère Hyot (1)
Affiliations
- CEA-LETI
- CEA-IRIG
Background
AlGaN LEDs have been greatly researched these last years for the development of ultraviolet emitters used in various applications, including disinfection [1]. For that purpose, we need AlGaN alloys with high Al content and high active dopant concentrations. In this presentation, we focus on the effect of Al percentage and n-type doping (Ge and Si) on our structures. Indeed, silicon is a widely used n-type dopant but it contributes to edge-type dislocation climb and induces tensile stress. The interest of Ge is that it introduces less strain than silicon in the AlGaN lattice. Moreover, the distribution and activity of dopants and the location of the electrical junction are very important parameters in the design and optimisation of LED structures. These properties can be measured using off-axis electron holography, which records both the amplitude and the electron phase from an interference pattern. Since changes in phase of the electron beam are directly linked to the electrostatic potential, this allows us to measure these properties with a nanometer resolution.
Methods
In this study, we have grown several Ge-doped AlGaN samples with varying AlN percentages and dopant concentrations, we also choose to grow a single sample with 5 50-nm-thick AlGaN layers separated by 5nm-thick GaN layers to evacuate the charges, on a GaN buffer. Each AlGaN layer had increasing AlN concentrations, ranging from 10% to 50%, to specifically analyze the effect of AlN percentage. Lamellas for transmission electron microscopy (TEM) studies were prepared by in-situ lift-out focused ion beam (FIB) milling in an FEI Strata 400 from samples grown by plasma-assisted molecular beam epitaxy (PAMBE). Each lamella contained areas with different thicknesses to obtain information about the samples such as the inactive thickness and to determine the optimal experimental conditions. Electron holography and electron dispersive x-ray (EDX) spectroscopy were then performed using a double-corrected FEI Titan Ultimate TEM operating at 200 kV, equipped with a Gatan One View 4k camera.
Results
Using electron holography, we were able to map the electrostatic potential within the specimens. This required high-quality sample preparation and optimized experimental parameters during TEM experiments [2]. Measurement conditions for AlGaN are particularly challenging due to dynamical diffraction and the tendency of wide-bandgap materials to charge under electron beam exposure. Additionally, the total potential measured in electron holography depends on strong polarization fields, the mean inner potential (MIP) of our material, the built in potential (resulting from differences in electron affinity), and additional contributions from dopants. To obtain reliable measurements, we can first deposit a thin carbon layer on the specimen to provide a conductive channel to evacuate the charge. Then, the theorical MIP in the AlGaN was calculated from DFT simulations, which provides values of 15,9 V for bulk AlN and 16,9 V for bulk GaN [3]. Through linear interpolation, we estimated the theoretical MIP of each AlGaN layer based on its composition. Figure 1(a) shows a FIB lamella with several steps of different thicknesses. In order to measure accurately the thicknesses of these steps, we employed the Convergent Beam Electron Diffraction (CBED) method [4]. Figure 1(b) shows an experimental CBED pattern compared to 1(c) a simulation for a 374nm thick specimen using the JEMS software. We also analyzed our sample using complementary characterization techniques, including STEM, EDX, and strain mapping (Figure 1(d)). However, the primary focus remains on electron holography. Indeed, in addition to the thickness, we also have to determine the phase. First, we examined the amplitude image of the specimen. If the amplitude image exhibits homogenous contrast, indicating reduced diffraction contrast, then the corresponding phase image (Figure 1(e)) can be reliably interpreted. From this image, we extracted the phase difference between two steps and then determined and compared electrostatic potentials across the AlGaN layer using the equation : ΔՓ0(x,y)=CE.V(x,y).t(x,y) where CE = 7,289 × 10−3 rad.V−1.nm−1 for 200kV electrons. Figure 1(f) shows the potential profile across our structure. We then compared the potential evolution with other structural characteristics, such as Al content and strain. Figure 1 (g) displays the profile we obtained from the EDX analysis and provides us precise information on Al atomic percentage, while Figure 1(h) illustrates the profile we obtained from the strain mapping and give us information on the stress and relaxation along the growth axis.
Conclusion
By optimizing experimental conditions, we will be able to show the effect of Al content on the activity of the dopants in the AlGaN layers. With this study, we also show that the off-axis electron holography is a useful method to measure the electrostatic potential with a nanometer resolution, and thus is a valuable tool to optimize structures grown for the development of UV LEDs. Additionally, we couple holography with EDX and strain mapping to study the physical and chemical influences on our structure as well as refining the values expected for the MIP.
This research was performed on the Platform for Nanocharacterisation (PFNC) at the CEA-LETI of Grenoble and was funded by the French National Research Agency (Grant ANR-22-CE09-0024).
Graphic
References
[1] H. Amano, et al., J. Phys. D: Appl. Phys. 53, 503001 (2020).
[2] M. R. McCartney et al., Appl. Phys. Lett. 76, 3055 (2000).
[3] M. Schowalter, Appl. Phys. Lett. 88, 232108. (2006).
[4] D.B Williams and C.B. Carter, Transmission Electron Microscopy; Springer US, (2009).
Keywords
Electron holography, AlGaN, n-type doping, electostatic potential